Ugarit is a backup/archival system based around content-addressible
storage.
News
Development priorities are: Performance, better error handling, and
fixing bugs! After I've cleaned house a little, I'll be focussing on
replicated backend storage (ticket [f1f2ce8cdc]), as I now have a
cluster of storage devices at home.
2015-06-12: Version 2.0 is released, containing rudimentary archive
mode, plus many minor improvements! See the release notes at the
bottom for more details.
2014-11-02: Chicken itself has gained
[http://code.call-cc.org/cgi-bin/gitweb.cgi?p=chicken-core.git;a=commit;h=a0ce0b4cb4155754c1a304c0d8b15276b11b8cd2|significantly
faster byte-vector I/O]. This is only on the trunk at the time of
writing; I look forward to it being in a formal release, as it sped up
Ugarit snapshot benchmarks (dumping a 256MiB file into an sqlite
backend) by a factor of twenty-something.
2014-02-21: User [http://rmm.meta.ph/|Rommel Martinez] has written
[http://rmm.meta.ph/blog/2014/02/21/an-introduction-to-ugarit/|An introduction to Ugarit]!
About Ugarit
What's content-addressible storage?
Traditional backup systems work by storing copies of your files
somewhere. Perhaps they go onto tapes, or perhaps they're in archive
files written to disk. They will either be full dumps, containing a
complete copy of your files, or incrementals or differentials, which
only contain files that have been modified since some point. This
saves making repeated copies of unchanging files, but it means that to
do a full restore, you need to start by extracting the last full dump
then applying one or more incrementials, or the latest differential,
to get the latest state.
Not only do differentials and incrementals let you save space, they
also give you a history - you can restore up to a previous point in
time, which is invaluable if the file you want to restore was deleted
a few backup cycles ago!
This technology was developed when the best storage technology for
backups was magnetic tape, because each dump is written sequentially
(and restores are largely sequential, unless you're skipping bits to
pull out specific files).
However, these days, random-access media such as magnetic disks and
SSDs are cheap enough to compete with magnetic tape for long-term bulk
storage (especially when one considers the cost of a tape drive or
two). And having fast random access means we can take advantage of
different storage techniques.
A content-addressible store is a key-value store, except that the keys
are always computed from the values. When a given object is stored, it
is hashed, and the hash used as the key. This means you can never
store the same object twice; the second time you'll get the same hash,
see the object is already present, and re-use the existing
copy. Therefore, you get deduplication of your data for free.
But, I hear you ask, how do you find things again, if you can't choose
the keys?
When an object is stored, you need to record the key so you can find
it again later. In Ugarit, everything is stored in a tree-like
directory structure. Files are uploaded and their hashes obtained, and
then a directory object is constructed containing a list of the files
in the directory, and listing the key of the Ugarit objects storing
the contents of each file. This directory object itself has a hash,
which is stored inside the directory entry in the parent directory,
and so on up to the root. The root of a tree stored in a Ugarit vault
has no parent directory to contain it, so at that point, we store the
key of the root in a named "tag" that we can look up by name when we
want it.
Therefore, everything in a Ugarit vault can be found by starting with
a named tag and retrieving the object whose key it contains, then
finding keys inside that object and looking up the objects they refer
to, until we find the object we want.
When you use Ugarit to back up your filesystem, it uploads a complete
snapshot of every file in the filesystem, like a full dump. But
because the vault is content-addressed, it automatically avoids
uploading anything it already has a copy of, so all we upload is an
incremental dump - but in the vault, it looks like a full dump, and so
can be restored on its own without having to restore a chain of incrementals.
Also, the same storage can be shared between multiple systems that all
back up to it - and the incremental upload algorithm will mean that
any files shared between the servers will only need to be uploaded
once. If you back up a complete server, than go and back up another
that is running the same distribution, then all the files in /bin
and so on that are already in the storage will not need to be backed
up again; the system will automatically spot that they're already
there, and not upload them again.
As well as storing backups of filesystems, Ugarit can also be used as
the primary storage for read-only files, such as music and photos. The
principle is exactly the same; the only difference is in how the files
are organised - rather than as a directory structure, the files are
referenced from metadata objects that specify information about the
file (so it can be found) and a reference to the contents. Sets of
metadata objects are pointed to by tags as well, so they can also be
found.
So what's that mean in practice?
Backups
You can run Ugarit to back up any number of filesystems to a shared
storage area (known as a vault, and on every backup, Ugarit
will only upload files or parts of files that aren't already in the
vault - be they from the previous snapshot, earlier snapshots,
snapshot of entirely unrelated filesystems, etc. Every time you do a
snapshot, Ugarit builds an entire complete directory tree of the
snapshot in the vault - but reusing any parts of files, files, or
entire directories that already exist anywhere in the vault, and
only uploading what doesn't already exist.
The support for parts of files means that, in many cases, gigantic
files like database tables and virtual disks for virtual machines will
not need to be uploaded entirely every time they change, as the
changed sections will be identified and uploaded.
Because a complete directory tree exists in the vault for any
snapshot, the extraction algorithm is incredibly simple - and,
therefore, incredibly reliable and fast. Simple, reliable, and fast
are just what you need when you're trying to reconstruct the
filesystem of a live server.
Also, it means that you can do lots of small snapshots. If you run a
snapshot every hour, then only a megabyte or two might have changed in
your filesystem, so you only upload a megabyte or two - yet you end up
with a complete history of your filesystem at hourly intervals in the
vault.
Conventional backup systems usually either store a full backup then
incrementals to their archives, meaning that doing a restore involves
reading the full backup then reading every incremental since and
applying them - so to do a restore, you have to download *every
version* of the filesystem you've ever uploaded, or you have to do
periodic full backups (even though most of your filesystem won't have
changed since the last full backup) to reduce the number of
incrementals required for a restore. Better results are had from
systems that use a special backup server to look after the archive
storage, which accept incremental backups and apply them to the
snapshot they keep in order to maintain a most-recent snapshot that
can be downloaded in a single run; but they then restrict you to using
dedicated servers as your archive stores, ruling out cheaply scalable
solutions like Amazon S3, or just backing up to a removable USB or
eSATA disk you attach to your system whenever you do a backup. And
dedicated backup servers are complex pieces of software; can you rely
on something complex for the fundamental foundation of your data
security system?
Archives
You can also use Ugarit as the primary storage for read-only
files. You do this by creating an archive in the vault, and importing
batches of files into it along with their metadata (arbitrary
attributes, such as "author", "creation date" or "subject").
Just as you can keep snapshots of multiple systems in a Ugarit vault,
you can also keep multiple separate archives, each identified by a
named tag.
However, as it's all within the same vault, the usual de-duplication
rules apply. The same file may be in multiple archives, with different
metadata in each, as the file contents and metadata are stored
separately (and associated only within the context of each
archive). And, of course, the same file may appear in snapshots and in
archives; perhaps a file was originally downloaded into your home
directory, where it was backed up into Ugarit snapshots, and then you
imported it into your archive. The archive import would not have had
to re-upload the file, as its contents would have already been found
in the vault, so all that needs to be uploaded is the metadata.
Although we have mainly spoken of storing files in archives, the
objects in archives can be files or directories full of files, as
well. This is useful for storing MacOS-style files that are actually
directories, or for archiving things like completed projects for
clients, which can be entire directory structures.
System Requirements
Ugarit should run on any POSIX-compliant system that can run
[http://www.call-with-current-continuation.org/|Chicken Scheme]. It
stores and restores all the file attributes reported by the stat
system call - POSIX mode permissions, UID, GID, mtime, and optionally
atime and ctime (although the ctime cannot be restored due to POSIX
restrictions). Ugarit will store files, directories, device and
character special files, symlinks, and FIFOs.
Support for extended filesystem attributes - ACLs, alternative
streams, forks and other metadata - is possible, due to the extensible
directory entry format; support for such metadata will be added as
required.
Currently, only local filesystem-based vault storage backends are
complete: these are suitable for backing up to a removable hard disk
or a filesystem shared via NFS or other protocols. However, the
backend can be accessed via an SSH tunnel, so a remote server you are
able to install Ugarit on to run the backends can be used as a remote
vault.
However, the next backend to be implemented will be one for Amazon S3,
and an SFTP backend for storing vaults anywhere you can ssh
to. Other backends will be implemented on demand; a vault can, in
principle, be stored on anything that can store files by name, report
on whether a file already exists, and efficiently download a file by
name. This rules out magnetic tapes due to their requirement for
sequential access.
Although we need to trust that a backend won't lose data (for now), we
don't need to trust the backend not to snoop on us, as Ugarit
optionally encrypts everything sent to the vault.
Terminology
A Ugarit backend is the software module that handles backend
storage. An actual storage area - managed by a backend - is called a
storage, and is used to implement a vault; currently, every storage is
a valid vault, but the planned future introduction of a distributed
storage backend will enable multiple storages (which are not,
themselves, valid vaults as they only contain some subset of the
information required) to be combined into an aggregrate storage, which
then holds the actual vault. Note that the contents of a storage is
purely a set of blocks, and a series of named tags containing
references to them; the storage does not know the details of
encryption and hashing, so cannot make any sense of its contents.
For example, if you use the recommended "splitlog" filesystem backend,
your vault might be /mnt/bigdisk on the server
prometheus. The backend (which is compiled along with the
other filesystem backends in the backend-fs binary) must
be installed on prometheus, and Ugarit clients all over
the place may then use it via ssh to prometheus. However,
even with the filesystem backends, the actual storage might not be on
prometheus where the backend runs -
/mnt/bigdisk might be an NFS mount, or a mount from a
storage-area network. This ability to delegate via SSH is particularly
useful with the "cache" backend, which reduces latency by storing a
cache of what blocks exist in a backend, thereby making it quicker to
identify already-stored files; a cluster of servers all sharing the
same vault might all use SSH tunnels to access an instance of the
"cache" backend on one of them (using some local disk to store the
cache), which proxies the actual vault storage to a vault on the other
end of a high-latency Internet link, again via an SSH tunnel.
A vault is where Ugarit stores backups (as chains of snapshots) and
archives (as chains of archive imports). Backups and archives are
identified by tags, which are the top-level named entry points into a
vault. A vault is based on top of a storage, along with a choice of
hash function, compression algorithm, and encryption that are used to
map the logical world of snapshots and archive imports into the
physical world of blocks stored in the storage.
A snapshot is a copy of a filesystem tree in the vault, with a header
block that gives some metadata about it. A backup consists of a number
of snapshots of a given filesystem.
An archive import is a set of filesystem trees, each along with
metadata about it. Whereas a backup is organised around a series of
timed snapshots, an archive is organised around the metadata; the
filesystem trees in the archive are identified by their properties.
So what, exactly, is in a vault?
A Ugarit vault contains a load of blocks, each up to a maximum size
(usually 1MiB, although other backends might impose smaller
limits). Each block is identified by the hash of its contents; this is
how Ugarit avoids ever uploading the same data twice, by checking to
see if the data to be uploaded already exists in the vault by
looking up the hash. The contents of the blocks are compressed and
then encrypted before upload.
Every file uploaded is, unless it's small enough to fit in a single
block, chopped into blocks, and each block uploaded. This way, the
entire contents of your filesystem can be uploaded - or, at least,
only the parts of it that aren't already there! The blocks are then
tied together to create a snapshot by uploading blocks full of the
hashes of the data blocks, and directory blocks are uploaded listing
the names and attributes of files in directories, along with the
hashes of the blocks that contain the files' contents. Even the blocks
that contain lists of hashes of other blocks are subject to checking
for pre-existence in the vault; if only a few MiB of your
hundred-GiB filesystem has changed, then even the index blocks and
directory blocks are re-used from previous snapshots.
Once uploaded, a block in the vault is never again changed. After all,
if its contents changed, its hash would change, so it would no longer
be the same block! However, every block has a reference count,
tracking the number of index blocks that refer to it. This means that
the vault knows which blocks are shared between multiple snapshots (or
shared *within* a snapshot - if a filesystem has more than one copy of
the same file, still only one copy is uploaded), so that if a given
snapshot is deleted, then the blocks that only that snapshot is using
can be deleted to free up space, without corrupting other snapshots by
deleting blocks they share. Keep in mind, however, that not all
storage backends may support this - there are certain advantages to
being an append-only vault. For a start, you can't delete something by
accident! The supplied fs and sqlite backends support deletion, while
the splitlog backend does not yet. However, the actual snapshot
deletion command in the user interface hasn't been implemented yet
either, so it's a moot point for now...
Finally, the vault contains objects called tags. Unlike the blocks,
the tags' contents can change, and they have meaningful names rather
than being identified by hash. Tags identify the top-level blocks of
snapshots within the system, from which (by following the chain of
hashes down through the index blocks) the entire contents of a
snapshot may be found. Unless you happen to have recorded the hash of
a snapshot somewhere, the tags are where you find snapshots from when
you want to do a restore.
Whenever a snapshot is taken, as soon as Ugarit has uploaded all the
files, directories, and index blocks required, it looks up the tag you
have identified as the target of the snapshot. If the tag already
exists, then the snapshot it currently points to is recorded in the
new snapshot as the "previous snapshot"; then the snapshot header
containing the previous snapshot hash, along with the date and time
and any comments you provide for the snapshot, and is uploaded (as
another block, identified by its hash). The tag is then updated to
point to the new snapshot.
This way, each tag actually identifies a chronological chain of
snapshots. Normally, you would use a tag to identify a filesystem
being backed up; you'd keep snapshotting the filesystem to the same
tag, resulting in all the snapshots of that filesystem hanging from
the tag. But if you wanted to remember any particular snapshot
(perhaps if it's the snapshot you take before a big upgrade or other
risky operation), you can duplicate the tag, in effect 'forking' the
chain of snapshots much like a branch in a version control system.
Archive imports cause the creation of one or more archive metadata
blocks, each of which lists the hashes of files or filesystem trees in
the archive, along with their metadata. Each import then has a single
archive import block pointing to the sequence of metadata blocks, and
pointing to the previous archive import block in that archive. The
same filesystem tree can be imported more than once to the same
archive, and the "latest" metadata always wins.
Generally, you should create lots of small archives for different
categories of things - such as one for music, one for photos, and so
on. You might well create separate archives for the music collections
of different people in your household, unless they overlap, and
another for Christmas music so it doesn't crop up in random shuffle
play! It's easy to merge archives if you over-compartmentalise them,
but harder to split an archive if you find it too cluttered with
unrelated things.
I've spoken of archive imports, and backup snapshots, each having a
"previous" reference to the last import or snapshot in the chain, but
it's actually more complex than that: they have an arbitrary list of
zero or more previous objects. As such, it's possible for several
imports or snapshots to have the same "previous", known as a "fork",
and it's possible to have an import or snapshot that merges multiple
previous ones.
Forking is handy if you want to basically duplicate an archive,
creating two new archives with the same contents to begin with, but
each then capable of diverging thereafter. You might do this to keep
the state of an archive before doing a bit import, so you can go back
to the original state if you regret the import, for instance.
Forking a backup tag is a more unusual operation, but also
useful. Perhaps you have a server running many stateful services, and
the hardware becomes overloaded, so you clone the basic setup onto
another server, and run half of the services on the original and half
on the new one; if you fork the backup tag of the original server to
create a backup tag for the new server, then both servers' snapshot
history will share the original shared state.
Merging is most useful for archives; you might merge several archives
into one, as mentioned.
And, of course, you can merge backup tags, as well. If your earlier
splitting of one server into two doesn't work out (perhaps your
workload reduces, or you can now afford a single, more powerful,
server to handle everything in one place), you might rsync back the
service state from the two servers onto the new server, so it's all
merged in the new server's filesystem. To preserve this in the
snapshot history, you can merge the two backup tags of the two servers
to create a backup tag for the single new server, which will
accurately reflect the history of the filesystem.
Also, tags might fork by accident - I plan to introduce a distributed
storage backend, which will replicate blocks and tags across multiple
storages to create a single virtual storage to build a vault on top
of; in the event of the network of actual storages suffering a
failure, it may be that snapshots and imports are only applied to some
of the storages - and then subsequent snapshots and imports only get
applied to some other subset of the storages. When the network is
repaired and all the storages are again visible, they will have
diverged, inconsistent, states for their tags, and the distributed
storage system will resolve the situation by keeping the majority
state as the state of the tag on all the backends, but preserving any
other states by creating new tags, with the original name plus a
suffix. These can then be merged to "heal" the conflict.
Using Ugarit
Installation
Install [http://www.call-with-current-continuation.org/|Chicken Scheme] using their [http://wiki.call-cc.org/man/4/Getting%20started|installation instructions].
Ugarit can then be installed by typing (as root):
chicken-install ugarit
See the [http://wiki.call-cc.org/manual/Extensions#chicken-install-reference|chicken-install manual] for details if you have any trouble, or wish to install into your home directory.
Setting up a vault
Firstly, you need to know the vault identifier for the place you'll
be storing your vaults. This depends on your backend. The vault
identifier is actually the command line used to invoke the backend for
a particular vault; communication with the vault is via standard
input and output, which is how it's easy to tunnel via ssh.
Local filesystem backends
These backends use the local filesystem to store the vaults. Of
course, the "local filesystem" on a given server might be an NFS mount
or mounted from a storage-area network.
Logfile backend
The logfile backend works much like the original Venti system. It's
append-only - you won't be able to delete old snapshots from a logfile
vault, even when I implement deletion. It stores the vault in two
sets of files; one is a log of data blocks, split at a specified
maximum size, and the other is the metadata: an sqlite database used
to track the location of blocks in the log files, the contents of
tags, and a count of the logs so a filename can be chosen for a new one.
To set up a new logfile vault, just choose where to put the two
parts. It would be nice to put the metadata file on a different
physical disk to the logs directory, to reduce seeking. If you only
have one disk, you can put the metadata file in the log directory
("metadata" is a good name).
You can then refer to it using the following vault identifier:
"backend-fs splitlog ...log directory... ...metadata file..."
SQLite backend
The sqlite backend works a bit like a
[http://www.fossil-scm.org/|Fossil] repository; the storage is
implemented as a single file, which is actually an SQLite database
containing blocks as blobs, along with tags and configuration data in
their own tables.
It supports unlinking objects, and the use of a single file to store
everything is convenient; but storing everything in a single file with
random access is slightly riskier than the simple structure of an
append-only log file; it is less tolerant of corruption, which can
easily render the entire storage unusable. Also, that one file can get
very large.
SQLite has internal limits on the size of a database, but they're
quite large - you'll probably hit a size limit at about 140
terabytes.
To set up an SQLite storage, just choose a place to put the file. I
usually use an extension of .vault; note that SQLite will
create additional temporary files alongside it with additional
extensions, too.
Then refer to it with the following vault identifier:
"backend-sqlite ...path to vault file..."
Filesystem backend
The filesystem backend creates vaults by storing each block or tag
in its own file, in a directory. To keep the objects-per-directory
count down, it'll split the files into subdirectories. Because of
this, it uses a stupendous number of inodes (more than the filesystem
being backed up). Only use it if you don't mind that; splitlog is much
more efficient.
To set up a new filesystem-backend vault, just create an empty
directory that Ugarit will have write access to when it runs. It will
probably run as root in order to be able to access the contents of
files that aren't world-readable (although that's up to you), so
unless you access your storage via ssh or sudo to use another user to
run the backend under, be careful of NFS mounts that have
maproot=nobody set!
You can then refer to it using the following vault identifier:
"backend-fs fs ...path to directory..."
Proxying backends
These backends wrap another vault identifier which the actual
storage task is delegated to, but add some value along the way.
SSH tunnelling
It's easy to access a vault stored on a remote server. The caveat
is that the backend then needs to be installed on the remote server!
Since vaults are accessed by running the supplied command, and then
talking to them via stdin and stdout, the vault identified needs
only be:
"ssh ...hostname... '...remote vault identifier...'"
Cache backend
The cache backend is used to cache a list of what blocks exist in the
proxied backend, so that it can answer queries as to the existance of
a block rapidly, even when the proxied backend is on the end of a
high-latency link (eg, the Internet). This should speed up snapshots,
as existing files are identified by asking the backend if the vault
already has them.
The cache backend works by storing the cache in a local sqlite
file. Given a place for it to store that file, usage is simple:
"backend-cache ...path to cachefile... '...proxied vault identifier...'"
The cache file will be automatically created if it doesn't already
exist, so make sure there's write access to the containing directory.
- WARNING - WARNING - WARNING - WARNING - WARNING - WARNING -
If you use a cache on a vault shared between servers, make sure
that you either:
* Never delete things from the vault
or
* Make sure all access to the vault is via the same cache
If a block is deleted from a vault, and a cache on that vault is
not aware of the deletion (as it did not go "through" the caching
proxy), then the cache will record that the block exists in the
vault when it does not. This will mean that if a snapshot is made
through the cache that would use that block, then it will be assumed
that the block already exists in the vault when it does
not. Therefore, the block will not be uploaded, and a dangling
reference will result!
Some setups which *are* safe:
* A single server using a vault via a cache, not sharing it with
anyone else.
* A pool of servers using a vault via the same cache.
* A pool of servers using a vault via one or more caches, and
maybe some not via the cache, where nothing is ever deleted from
the vault.
* A pool of servers using a vault via one cache, and maybe some
not via the cache, where deletions are only performed on servers
using the cache, so the cache is always aware.
Writing a ugarit.conf
ugarit.conf should look something like this:
(storage )
(hash tiger "")
[double-check]
[(compression [deflate|lzma])]
[(encryption aes )]
[(file-cache "")]
[(rule ...)]
The hash line chooses a hash algorithm. Currently Tiger-192
(tiger), SHA-256 (sha256), SHA-384
(sha384) and SHA-512 (sha512) are supported;
if you omit the line then Tiger will still be used, but it will be a
simple hash of the block with the block type appended, which reveals
to attackers what blocks you have (as the hash is of the unencrypted
block, and the hash is not encrypted). This is useful for development
and testing or for use with trusted vaults, but not advised for use
with vaults that attackers may snoop at. Providing a salt string
produces a hash function that hashes the block, the type of block, and
the salt string, producing hashes that attackers who can snoop the
vault cannot use to find known blocks (see the "Security model"
section below for more details).
I would recommend that you create a salt string from a secure entropy
source, such as:
dd if=/dev/random bs=1 count=64 | base64 -w 0
Whichever hash function you use, you will need to install the required
Chicken egg with one of the following commands:
chicken-install -s tiger-hash # for tiger
chicken-install -s sha2 # for the SHA hashes
double-check, if present, causes Ugarit to perform extra
internal consistency checks during backups, which will detect bugs but
may slow things down.
lzma is the recommended compression option for
low-bandwidth backends or when space is tight, but it's very slow to
compress; deflate or no compression at all are better for fast local
vaults. To have no compression at all, just remove the
(compression ...) line entirely. Likewise, to use
compression, you need to install a Chicken egg:
chicken-install -s z3 # for deflate
chicken-install -s lzma # for lzma
WARNING: The lzma egg is currently rather difficult to install, and
needs rewriting to fix this problem.
Likewise, the (encryption ...) line may be omitted to have no
encryption; the only currently supported algorithm is aes (in CBC
mode) with a key given in hex, as a passphrase (hashed to get a key),
or a passphrase read from the terminal on every run. The key may be
16, 24, or 32 bytes for 128-bit, 192-bit or 256-bit AES. To specify a
hex key, just supply it as a string, like so:
...for 256-bit AES.
Alternatively, you can provide a passphrase, and specify how large a
key you want it turned into, like so:
(encryption aes ([16|24|32] "We three kings of Orient are, one in a taxi one in a car, one on a scooter honking his hooter and smoking a fat cigar. Oh, star of wonder, star of light; star with royal dynamite"))
I would recommend that you generate a long passphrase from a secure
entropy source, such as:
dd if=/dev/random bs=1 count=64 | base64 -w 0
Finally, the extra-paranoid can request that Ugarit prompt for a
passphrase on every run and hash it into a key of the specified
length, like so:
(encryption aes ([16|24|32] prompt))
(note the lack of quotes around prompt, distinguishing it from a passphrase)
Please read the "Security model" section below for details on the
implications of different encryption setups.
Again, as it is an optional feature, to use encryption, you must
install the appropriate Chicken egg:
chicken-install -s aes
A file cache, if enabled, significantly speeds up subsequent snapshots
of a filesystem tree. The file cache is a file (which Ugarit will
create if it doesn't already exist) mapping filenames to
(mtime,size,hash) tuples; as it scans the filesystem, if it finds a
file in the cache and the mtime and size have not changed, it will
assume it is already stored under the specified hash. This saves it
from having to read the entire file to hash it and then check if the
hash is present in the vault. In other words, if only a few files
have changed since the last snapshot, then snapshotting a directory
tree becomes an O(N) operation, where N is the number of files, rather
than an O(M) operation, where M is the total size of files involved.
For example:
Be careful to put a set of parentheses around each configuration
entry. White space isn't significant, so feel free to indent things
and wrap them over lines if you want.
Keep copies of this file safe - you'll need it to do extractions!
Print a copy out and lock it in your fire safe! Ok, currently, you
might be able to recreate it if you remember where you put the
storage, but encryption keys and hash salts are harder to remember...
Your first backup
Think of a tag to identify the filesystem you're backing up. If it's
/home on the server gandalf, you might call it gandalf-home. If
it's the entire filesystem of the server bilbo, you might just call
it bilbo.
Then from your shell, run (as root):
# ugarit snapshot [-c] [-a]
For example, if we have a ugarit.conf in the current directory:
Specify the -c flag if you want to store ctimes in the vault;
since it's impossible to restore ctimes when extracting from an
vault, doing this is useful only for informational purposes, so it's
not done by default. Similarly, atimes aren't stored in the vault
unless you specify -a, because otherwise, there will be a lot of
directory blocks uploaded on every snapshot, as the atime of every
file will have been changed by the previous snapshot - so with -a
specified, on every snapshot, every directory in your filesystem will
be uploaded! Ugarit will happily restore atimes if they are found in
a vault; their storage is made optional simply because uploading
them is costly and rarely useful.
Exploring the vault
Now you have a backup, you can explore the contents of the
vault. This need not be done as root, as long as you can read
ugarit.conf; however, if you want to extract files, run it as root
so the uids and gids can be set.
$ ugarit explore ugarit.conf
This will put you into an interactive shell exploring a virtual
filesystem. The root directory contains an entry for every tag; if you
type ls you should see your tag listed, and within that
tag, you'll find a list of snapshots, in descending date order, with a
special entry current for the most recent
snapshot. Within a snapshot, you'll find the root directory of your
snapshot under contents, and the detailts of the snapshot itself in
propreties.sexpr, and will be able to cd into
subdirectories, and so on:
> ls
localhost-etc/
> cd localhost-etc
/localhost-etc> ls
current/
2015-06-12 22:49:34/
2015-06-12 22:49:25/
/localhost-etc> cd current
/localhost-etc/current> ls
log.sexpr
properties.sexpr
contents/
/localhost-etc/current> cat properties.sexpr
((previous . "a140e6dbe0a7a38f8b8c381323997c23e51a39e2593afb61")
(mtime . 1434102574.0)
(contents . "34eccf1f5141187e4209cfa354fdea749a0c3c1c4682ec86")
(stats (blocks-stored . 12)
(bytes-stored . 16889)
(blocks-skipped . 50)
(bytes-skipped . 6567341)
(file-cache-hits . 0)
(file-cache-bytes . 0))
(log . "b2a920f962c12848352f33cf32941e5313bcc5f209219c1a")
(hostname . "ahe")
(source-path . "/etc")
(notes)
(files . 112)
(size . 6563588))
/localhost-etc/current> cd contents
/localhost-etc/current/contents> ls
zoneinfo
vconsole.conf
udev/
tmpfiles.d/
systemd/
sysctl.d/
sudoers.tmp~
sudoers
subuid
subgid
static
ssl/
ssh/
shells
shadow-
shadow
services
samba/
rpc
resolvconf.conf
resolv.conf
-- Press q then enter to stop or enter for more...
q
/localhost-etc/current/contents> ls -ll resolv.conf
-rw-r--r-- 0 0 [2015-05-23 23:22:41] 78B/-: resolv.conf
key: #f
contents: "e33ea1394cd2a67fe6caab9af99f66a4a1cc50e8929d3550"
size: 78
ctime: 1432419761.0
As well as exploring around, you can also extract files or directories
(or entire snapshots) by using the get command. Ugarit
will do its best to restore the metadata of files, subject to the
rights of the user you run it as.
Type help to get help in the interactive shell.
The interactive shell supports command-line editing, history and tab
completion for your convenience.
Extracting things directly
As well as using the interactive explore mode, it is also possible to
directly extract something from the vault, given a path.
Given the sample vault from the previous example, it would be possible
to extract the README.txt file with the following
command:
As mentioned above, you can fork a tag, creating two tags that
refer to the same snapshot and its history but that can then have
their own subsequent history of snapshots applied to each
independently, with the following command:
$ ugarit fork
Merging tags
And you can also merge two or more tags into one. It's possible to
merge a bunch of tags to make an entirely new tag, or you can merge a
tag into an existing tag, by having the "output" tag also be one of
the "input" tags.
The command to do this is:
$ ugarit merge
For instance, to import your classical music collection into your main
musical collection, you might do:
To import some files into an archive, you must create a manifest file
listing them, and their metadata. The manifest can also list
metadata for the import as a whole, perhaps naming the source of the
files, or the reason for importing them.
The metadata for a file (or an import) is a series of named
properties. The value of a property can be any Scheme value, written
in Scheme syntax (with strings double-quoted unless they are to be
interpreted as symbols), but strings and numbers are the most useful
types.
You can use whatever names you like for properties in metadata, but
there are some that the system applies automatically, and an informal
standard of sorts, which is documented in [docs/archive-schema.wiki].
You can produce a manifest file by hand, or use the Ugarit Manifest
Maker to produce one for you. You do this by installing it like so:
$ chicken-install ugarit-manifest-maker
And then running it, giving it any number of file and directory names
on the command line. When given directories, it will recursively scan
them to find all the files contained therein and put them in the
manifest; it will not put directories in the manifest, although it is
perfectly legal for you to do so when writing a manifest by hand. This
is because the manifest maker can't do much useful analysis on a
directory to suggest default metadata for them (so there isn't much
point in using it), and it's far more useful for it to make it easy
for you to import a large number of files individually by referencing
the directory containing them.
The manifest is sent to standard output, so you need to redirect it to
a file, like so:
$ ugarit-manifest-maker ~/music > music.manifest
You can specify command-line options, as well. -e PATTERN
or --exclude=PATTERN introduces a glob pattern for files
to exclude from the manifest, and -D KEY=VALUE or
--define=KEY=VALUE provides a property to be added to
every file in the manifest (as opposed to an import property, that is
part of the metadata of the overall import). Note that
VALUE must be double-quoted if it's a string, as per
Scheme value syntax.
One might use this like so:
The manifest maker simplifies the writing of manifests for files, by
listing the files in manifest format along with useful metadata
extracted from the filename and the file itself. For supported file
types (currently, MP3 and OGG music files), it will even look inside
the file to extract metadata.
The manifest file it generates will contain lots of comments
mentioning things it couldn't automatically analyse (such as unknown
OGG/ID3 tags, or unknown types of files); and for metadata properties
it thinks might be relevant but can't automatically provide, it
suggests them with an empty property declaration, commented out. The
idea is that, after generating a manifest, you read it by hand in a
text editor to attempt to improve it.
The format of a manifest file
Manifest files have a relatively simple format. The are based on
Scheme s-expressions, so can contain comments. From any semicolon (not
in a string or otherwise quoted) to the end of the line is a comment,
and #; in front of something comments out that something.
Import metadata properties are specified like so:
(KEY = VALUE)
...where, as usual, VALUE must be double-quoted if it's a
string.
Files to import, with their metadata, are specified like so:
(object "PATH OF FILE TO IMPORT"
(KEY = VALUE)
(KEY = VALUE)...
)
The closing parenthesis need not be on a line of its own, it's
conventionally placed after the closing parenthesis of the final
property.
Ugarit, when importing the files in the manifest, will add the
following properties if they are not already specified:
import-path
The path the file was imported from
dc:format
A guess at the file's MIME type, based on the extension
mtime
The file's modification time (as the number of seconds since the
UNIX epoch)
ctime
The file's change time (as the number of seconds since the UNIX
epoch)
filename
The name of the file, stripped of any directory components, and
including the extension.
The following properties are placed in the import metadata,
automatically:
hostname
The hostname the import was performed on.
manifest-path
The path to the manifest file used for the import.
mtime
The time (in seconds since the UNIX epoch) at which the import was
committed.
stats
A Scheme alist of statistics about the import (number of
files/blocks uploaded, etc).
So, to wrap that all up, here's a sample import manifest file:
(notes = "A bunch of old CDs I've finally ripped")
(object "/home/alaric/newrip/track01.mp3"
(filename = "track01.mp3")
(dc:format = "audio/mpeg")
(dc:publisher = "Go! Beat Records")
(dc:created = "1994")
(dc:contributor = "Portishead")
(dc:subject = "Trip-Hop")
(superset:size = 1)
(superset:index = 1)
(set:title = "Dummy")
(set:size = 11)
(set:index = 1)
(dc:creator = "Portishead")
(dc:title = "Wandering Star")
(mtime = 1428962299.0)
(ctime = 1428962299.0)
(file-size = 4703055))
;;... and so on, for ten more MP3s on this CD, then several other CDs...
Actually importing a manifest
Well, when you finally have a manifest file, importing it is easy:
$ ugarit import
How do I change the metadata of an already-imported file?
That's easy; the "current" metadata of a file is the metadata of its
most recent. Just import the file again, in a new manifest, with new
metadata, and it will overwrite the old. However, the old metadata is
still preserved in the archive's history; tags forked from the archive
tag before the second import will still see the original state of the
archive, by design.
Exploring
Archives are visible in the explore interface. For instance, an import
of some music I did looks like this:
> ls
localhost-etc/ <tag>
archive-tag/ <tag>
> cd archive-tag
/archive-tag> ls
history/ <archive-history>
/archive-tag> cd history
/archive-tag/history> ls
2015-06-12 22:53:13/ <import>
/archive-tag/history> cd 2015-06-12 22:53:13
/archive-tag/history/2015-06-12 22:53:13> ls
log.sexpr <file>
properties.sexpr <inline>
manifest/ <import-manifest>
/archive-tag/history/2015-06-12 22:53:13> cat properties.sexpr
((stats (blocks-stored . 2046)
(bytes-stored . 1815317503)
(blocks-skipped . 9)
(bytes-skipped . 8388608)
(file-cache-hits . 0)
(file-cache-bytes . 0))
(log . "b2a920f962c12848352f33cf32941e5313bcc5f209219c1a")
(mtime . 1434135993.0)
(contents . "fcdd5b996914fdcac1e8a6cfbc67663e08f6eaf0cc952e21")
(hostname . "ahe")
(notes . "A bunch of music, imported as a demo")
(manifest-path . "/home/alaric/tmp/test.manifest"))
/archive-tag/history/2015-06-12 22:53:13> cd manifest
/archive-tag/history/2015-06-12 22:53:13/manifest> ls
1d4269099189234eefeb80b95370eaf280730cf4d591004d:03 The Lemon Song.mp3 <file>
7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382:04 Dazed and Confused.mp3 <file>
64092fa12c2800dda474b41e5ebe8c948f39a59ee91c120b:09 How Many More Times.mp3 <file>
1d79148d1e1e8947c50b44cf2d5690588787af328e82eeef:2-07 Going to California.mp3 <file>
e3685148d0d12213074a9fdb94a00e05282aeabe77fa60d5:1-01 You Shook Me.mp3 <file>
d73904f371af8d7ca2af1076881230f2dc1c2cf82416880a:03 Strangers.mp3 <file>
9c5a0efb7d397180a1e8d42356d8f04c6c26a83d3b05d34a:09 Uptight.mp3 <file>
01a069aec2e731e18fcdd4ecb0e424f346a2f0e16910f5e9:07 Numb.mp3 <file>
7ea1ab7fbd525c40e21d6dd25130e8c70289ad56c09375b0:08 She.mp3 <file>
009dacd8f3185b7caeb47050002e584ab86d08cf9e9aceec:1-03 Communication Breakdown.mp3 <file>
26d264d629e22709f664ed891741f690900d45cd4fd44326:1-03 Dazed and Confused.mp3 <file>
d879761195faf08e4e95a5a2398ea6eefb79920710bfeab6:1-10 Band Introduction _ How Many More Times.mp3 <file>
83244601db42677d110fc8522c6a3cbbc1f22966a779f876:06 All My Love.mp3 <file>
5eebee9a2ad79d04e4f69e9e2a92c4e0a8d5f21e670f89da:07 Tangerine.mp3 <file>
dd6f1203b5973ecd00d2c0cee18087030490230727591746:2-08 That's the Way.mp3 <file>
c0acea15aa27a6dd1bcaff1c13d4f3d741a40a46abeca3fc:04 The Crunge.mp3 <file>
ea7727ad07c6c82e5c9c7218ee1b059cd78264c131c1438d:1-02 I Can't Quit You Baby.mp3 <file>
10fda5f46b8f505ca965bcaf12252eedf5ab44514236f892:14 F.O.D..mp3 <file>
a99ca9af5a83bde1c676c388dc273051defa88756df26e95:1-03 Good Times Bad Times.mp3 <file>
b5d7cfe9808c7fc0dedbd656d44e4c56159cbd3c2ed963bb:1-15 Stairway to Heaven.mp3 <file>
79c87e3c49ffdac175c95aae071f63d3a9efdf2ddb84998c:08.Batmilk.ogg <file>
-- Press q then enter to stop or enter for more...
q
/archive-tag/history/2015-06-12 22:53:13/manifest> ls -ll 7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382:04 Dazed and Confused.mp3
-r-------- - - [2015-04-13 21:46:39] -/-: 7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382:04 Dazed and Confused.mp3
key: #f
contents: "7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382"
import-path: "/home/alaric/archive/sorted-music/Led Zeppelin/Led Zeppelin/04 Dazed and Confused.mp3"
filename: "04 Dazed and Confused.mp3"
dc:format: "audio/mpeg"
dc:publisher: "Atlantic"
dc:subject: "Classic Rock"
dc:title: "Dazed and Confused"
dc:creator: "Led Zeppelin"
dc:created: "1982"
dc:contributor: "Led Zeppelin"
set:title: "Led Zeppelin"
set:index: 4
set:size: 9
superset:index: 1
superset:size: 1
ctime: 1428957999.0
file-size: 15448903
Searching
However, the explore interface to an archive is far from pleasant. You
need to go to the correct import, and find your file by name, and then
identify it with a big long name composed of its hash and the original
filename to find its properties and extract.
I hope to add property-based searching to explore mode in future
(which is why you need to go into a history directory
within the archive directory, as other ways of exploring the archive
will appear alongside). This will be particularly useful when the
explore-mode virtual filesystem is mounted over 9P!
However, even that interface, being constrained to look like a
filesystem, will be limited. The ugarit command-line tool
provides a very powerful search interface that exposes the full power
of the archive metadata.
Metadata filters
Files (and directories) in an archive can be searched for using
"metadata filters", which are descriptions of what you're looking for
that the computer can understand. They are represented as Scheme
s-expressions, and can be made up of the following components:
#t
This filter matches everything. It's not very useful.
#f
This filter matches nothing. It's not very useful.
(and FILTER FILTER...)
This filter matches files for which all of the inner filters match.
(or FILTER FILTER...)
This filter matches files for which any of the inner filters match.
(not FILTER)
This filter matches files which do not match the inner filter.
(= ($ PROP) VALUE)
This filter matches files which have the given
PROPerty equal to that VALUE in their metadata.
(= key HASH)
This filter matches the file with the given hash.
(= ($import PROP) VALUE)
This filter matches files which have the given
PROPerty equal to that VALUE in the metadata
of the import that last imported them.
Searching an archive
For a start, you can search for files matching a given metadata filter
in a given archive. This is done with:
$ ugarit search
For instance, let's look for music by Led Zeppelin:
The result looks like the explore-mode view of an archive manifest,
listing the file's hash followed by its title and extension:
7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382:04 Dazed and Confused.mp3
834a1619a59835e0c27b22801e3c829b40be583dadd19770:2-08 No Quarter.mp3
9e8bc4954838bd9c671f275eb48595089257185750d63894:1-12 I Can't Quit You Baby.mp3
6742b3bebcdd9cae5ec5403c585935403fa74d16ed076cf2:02 Friends (1).mp3
07d161f4bd684e283f7f2cf26e0b732157a8e95ef66939c3:05 Carouselambra.mp3
[...]
What of all our lovely metadata? You can view that if you add the word
"verbose" to the end of the command line, which allows you to specify
alternate output formats:
Now the output looks like:
object a444ff6ef807b080b536155f58d246d633cab4a0eabef5bf
(ctime = 1428958660.0)
(dc:contributor = "Led Zeppelin")
(dc:created = "2008")
(dc:creator = "Led Zeppelin")
[... all the usual file properties omitted ...]
import a43f7a7268ee8b18381c20d7573add5dbf8781f81377279c
(stats = ((blocks-stored . 2046) (bytes-stored . 1815317503) (blocks-skipped . 9) (bytes-skipped . 8388608) (file-cache-hits . 0) (file-cache-bytes . 0)))
(log = "b2a920f962c12848352f33cf32941e5313bcc5f209219c1a")
[... all the usual import properties omitted ...]
object b4cadf48b2c07ccf0303fc4064b292cb222980b0d4223641
(ctime = 1428958673.0)
(dc:contributor = "Led Zeppelin")
(dc:created = "2008")
(dc:creator = "Led Zeppelin")
(dc:creator = "Jimmy Page/John Paul Jones/Robert Plant")
[...and so on...]
As you can see, it lists the hash of each file, its metadata, the hash
of the import that last imported it, and the metadata of that import.
That's quite verbose, so you'd probably be wanting to take that as
input to another program to do something nicer with it. But it's laid
out for human reading, not for machine parsing. Thankfully, we have
other formats for that, alist and
alist-with-imports.
Try this:
This outputs one Scheme s-expression list per match, the first element
of which is the hash as a string, the rest of which is an alist of properties:
("7cb253a4886b3e0051ea8cc0e78fb3a0160307a2c37c8382"
(ctime . 1428957999.0)
(dc:contributor . "Led Zeppelin")
(dc:created . "1982")
(dc:creator . "Led Zeppelin")
[... elided file properties ...]
(superset:index . 1)
(superset:size . 1))
("77c960d09eb21ed72e434ddcde0bd3781a4f3d6ee7a6eb66"
(ctime . 1428958981.0)
(dc:contributor . "Led Zeppelin")
[...]
This outputs one s-expression per list per match, with four
elements. The first is the key string, the second is an alist of file
properties, the third is the import's hash, and the last is an alist
containing the import's properties. It looks like:
("64fa08a0080aee6ef501c408fd44dfcc634cfcafd8006fc4"
((ctime . 1428958683.0)
(dc:contributor . "Led Zeppelin")
(dc:created . "2008")
(dc:creator . "Led Zeppelin")
[... elided file properties ...]
(superset:index . 1)
(superset:size . 1))
"a43f7a7268ee8b18381c20d7573add5dbf8781f81377279c"
((stats (blocks-stored . 2046)
(bytes-stored . 1815317503)
[... elided manifest properties ...]
(manifest-path . "test.manifest")))
("4cd56f916a63399b252976e842dcae0b87f058b5a60c93a4"
((ctime . 1428958437.0)
(dc:contributor . "Led Zeppelin")
[...]
And finally, you might just want to get the hashes of matching files
(which are particularly useful for extraction operations, which we'll
come to next). To do this, specify a format of "keys", which outputs
one line per match, containing just the hash:
ce6f6484337de772de9313038cb25d1b16e28028136cc291
6af5c664cbfa1acb22a377e97aee35d94c0fc003d239dd0c
92e91e79b384478b5aab31bf1b2ff9e25e7e2c4b48575185
6ddb9a41d4968468a904f05ecf7e0e73d2c7c7ad76bc394b
a074dddcef67cd93d92c6ffce845894aa56594674023f6e1
4f65f735bbb00a6fda4bc887b370b3160f55e5e07ec37ffa
97cc8b8ba70c39387fc08ef62311b751aea4340d636eb421
72358dbe3eb60da42eadcf6de325b2a6686f4e17ea41fa60
[...]
However, to write filter expressions, you need to know what properties
you have available to search on. You might remember, or go for
standard properties, or look at existing files in verbose mode to find
some; but you can also just ask Ugarit what properties it has in an
archive, like so:
$ ugarit search-props
You can even ask what properties are available for files matching an
existing filter:
$ ugarit search-props
This is useful if you're interested in further narrowing down a
filter, and so only care about properties that files already matching
that filter have.
For a bunch of music files imported with the Ugarit Manifest Maker,
you can expect to see something like this:
ctime
dc:contributor
dc:created
dc:creator
dc:format
dc:publisher
dc:subject
dc:title
file-size
filename
import-path
mtime
set:index
set:size
set:title
superset:index
superset:size
Now you know what properties to search, next you'll be wanting to know
what values to look for. Again, Ugarit has a command to query the
available values of any given property:
$ ugarit search-values
And you can limit that just to files matching a given filter:
$ ugarit search-values
The resulting list of values is ordered by popularity, so the most
widely-used values will be listed first. Let's see what genres of
music were in my sample of music files I imported:
The result is:
Classic Rock
Alternative & Punk
Electronic
Trip-Hop
Ok, let's now use a filter to find out what artists
(dc:creator) I have that made Trip-Hop music (what even
IS that?):
The result is:
Portishead
Ah, OK, now I know what "Trip-Hop" is.
Extracting
All this searching is lovely, but what it gets us, in the end, is a
bunch of file hashes. Perhaps we might want to actually play some
music, or look at a photo, or something. To do that, we need to
extract from the archive.
We've already seen the contents of an archive in the explore mode
virtual filesystem, so we could go into the archive history, find the
import, go into the manifest, pick the file out there, and use
get to extract it, but that would be yucky. Thankfully,
we have a command-line interface to get things from archives, in one
of two ways.
Firstly, we can extract a file (or a directory tree) from an archive,
out into the local filesystem:
$ ugarit archive-extract
The "target" is the name to give it in the local filesystem. We could
pull out that Led Zeppelin song from our search results above, like so:
We now have a foo.mp3 file in the current directory.
However, sometimes it would be nicer to have it streamed to standard
output, which can be done like so:
Each backend offers a number of administrative commands for
administering the storage underlying vaults. These are accessible via
the ugarit-storage-admin command line interface.
To use it, run it with the following command:
$ ugarit-storage-admin ''
The available commands differ between backends, but all backends
support the info and help commands, which
give basic information about the vault, and list all available
commands, respectively. Some offer a stats command that
examines the vault state to give interesting statistics, but which may
be a time-consuming operation.
Administering splitlog storages
The splitlog backend offers a wide selection of administrative
commands. See the help command on a splitlog vault for
details. The following commands are available:
help
List the available commands.
info
List some basic information about the storage.
stats
Examine the metadata to provide overall statistics about the
archive. This may be a time-consuming operation on large
storages.
set-block-size! BYTES
Sets the block size to the given number of bytes. This will affect
new blocks written to the storage, and leave existing blocks
untouched, even if they are larger than the new block size.
set-max-logfile-size! BYTES
Sets the size at which a log file is finished and a new one
started (likewise, existing log files will be untouched; this will
only affect new log files)
set-commit-interval! UPDATES
Sets the frequency of automatic synching of the storage
state to disk. Lowering this harms performance when writing to the
storage, but decreases the number of in-progress block writes that
can fail in a crash.
write-protect!
Disables updating of the storage.
write-unprotect!
Re-enables updating of the storage.
reindex!
Reindex the storage, rebuilding the block and tag state from the
contents of the log. If the metadata file is damaged or lost,
reindexing can rebuild it (although any configuration changes made
via other admin commands will need manually repeating as they are
not logged).
Administering sqlite storages
The sqlite backend has a similar administrative interface to the
splitlog backend, except that it does not have log files, so lacks the
set-max-logfile-size! and reindex! commands.
Administering cache storages
The cache backend provides a minimalistic interface:
help
List the available commands.
info
List some basic information about the storage.
stats
Report on how many entries are in the cache.
clear!
Clears the cache, dropping all the entries in it.
.ugarit files
By default, Ugarit will vault everything it finds in the filesystem
tree you tell it to snapshot. However, this might not always be
desired; so we provide the facility to override this with .ugarit
files, or global rules in your .conf file.
Note: The syntax of these files is provisional, as I want to
experiment with usability, as the current syntax is ugly. So please
don't be surprised if the format changes in incompatible ways in
subsequent versions!
In quick summary, if you want to ignore all files or directories
matching a glob in the current directory and below, put the following
in a .ugarit file in that directory:
(* (glob "*~") exclude)
You can write quite complex expressions as well as just globs. The
full set of rules is:
* (glob "pattern") matches files and directories whose names
match the glob pattern
* (name "name") matches files and directories with exactly that
name (useful for files called *...)
* (modified-within number seconds) matches files and
directories modified within the given number of seconds
* (modified-within number minutes) matches files and
directories modified within the given number of minutes
* (modified-within number hours) matches files and directories
modified within the given number of hours
* (modified-within number days) matches files and directories
modified within the given number of days
* (not rule) matches files and directories that do not match
the given rule
* (and rulerule...) matches files and directories that match
all the given rules
* (or rulerule...) matches files and directories that match
any of the given rules
Also, you can override a previous exclusion with an explicit include
in a lower-level directory:
(* (glob "*~") include)
You can bind rules to specific directories, rather than to "this
directory and all beneath it", by specifying an absolute or relative
path instead of the `*`:
("/etc" (name "passwd") exclude)
If you use a relative path, it's taken relative to the directory of
the .ugarit file.
You can also put some rules in your .conf file, although relative
paths are illegal there, by adding lines of this form to the file:
(rule * (glob "*~") exclude)
Questions and Answers
What happens if a snapshot is interrupted?
Nothing! Whatever blocks have been uploaded will be uploaded, but the
snapshot is only added to the tag once the entire filesystem has been
snapshotted. So just start the snapshot again. Any files that have
already be uploaded will then not need to be uploaded again, so the
second snapshot should proceed quickly to the point where it failed
before, and continue from there.
Unless the vault ends up with a partially-uploaded corrupted block
due to being interrupted during upload, you'll be fine. The filesystem
backend has been written to avoid this by writing the block to a file
with the wrong name, then renaming it to the correct name when it's
entirely uploaded.
Actually, there is *one* caveat: blocks that were uploaded, but never
make it into a finished snapshot, will be marked as "referenced" but
there's no snapshot to delete to un-reference them, so they'll never
be removed when you delete snapshots. (Not that snapshot deletion is
implemented yet, mind). If this becomes a problem for people, we could
write a "garbage collect" tool that regenerates the reference counts
in a vault, leading to unused blocks (with a zero refcount) being
unlinked.
Should I share a single large vault between all my filesystems?
I think so. Using a single large vault means that blocks shared
between servers - eg, software installed from packages and that sort
of thing - will only ever need to be uploaded once, saving storage
space and upload bandwidth. However, do not share a vault between
servers that do not mutually trust each other, as they can all update
the same tags, so can meddle with each other's snapshots - and read
each other's snapshots.
CAVEAT
It's not currently safe to have multiple concurrent snapshots to the
same split log backend; this will soon be fixed, however.
Security model
I have designed and implemented Ugarit to be able to handle cases
where the actual vault storage is not entirely trusted.
However, security involves tradeoffs, and Ugarit is configurable in
ways that affect its resistance to different kinds of attacks. Here I
will list different kinds of attack and explain how Ugarit can deal
with them, and how you need to configure it to gain that
protection.
Vault snoopers
This might be somebody who can intercept Ugarit's communication with
the vault at any point, or who can read the vault itself at their
leisure.
Ugarit's splitlog backend creates files with "rw-------" permissions
out of the box to try and prevent this. This is a pain for people who
want to share vaults between UIDs, but we can add a configuration
option to override this if that becomes a problem.
Reading your data
If you enable encryption, then all the blocks sent to the vault are
encrypted using a secret key stored in your Ugarit configuration
file. As long as that configuration file is kept safe, and the AES
algorithm is secure, then attackers who can snoop the vault cannot
decode your data blocks. Enabling compression will also help, as the
blocks are compressed before encrypting, which is thought to make
cryptographic analysis harder.
Recommendations: Use compression and encryption when there is a risk
of vault snooping. Keep your Ugarit configuration file safe using
UNIX file permissions (make it readable only by root), and maybe store
it on a removable device that's only plugged in when
required. Alternatively, use the "prompt" passphrase option, and be
prompted for a passphrase every time you run Ugarit, so it isn't
stored on disk anywhere.
Looking for known hashes
A block is identified by the hash of its content (before compression
and encryption). If an attacker was trying to find people who own a
particular file (perhaps a piece of subversive literature), they could
search Ugarit vaults for its hash.
However, Ugarit has the option to "key" the hash with a "salt" stored
in the Ugarit configuration file. This means that the hashes used are
actually a hash of the block's contents *and* the salt you supply. If
you do this with a random salt that you keep secret, then attackers
can't check your vault for known content just by comparing the hashes.
Recommendations: Provide a secret string to your hash function in your
Ugarit configuration file. Keep the Ugarit configuration file safe, as
per the advice in the previous point.
Vault modifiers
These folks can modify Ugarit's writes into the vault, its reads
back from the vault, or can modify the vault itself at their leisure.
Modifying an encrypted block without knowing the encryption key can at
worst be a denial of service, corrupting the block in an unknown
way. An attacker who knows the encryption key could replace a block
with valid-seeming but incorrect content. In the worst case, this
could exploit a bug in the decompression engine, causing a crash or
even an exploit of the Ugarit process itself (thereby gaining the
powers of a process inspector, as documented below). We can but hope
that the decompression engine is robust. Exploits of the decryption
engine, or other parts of Ugarit, are less likely due to the nature of
the operations performed upon them.
However, if a block is modified, then when Ugarit reads it back, the
hash will no longer match the hash Ugarit requested, which will be
detected and an error reported. The hash is checked after
decryption and decompression, so this check does not protect us
against exploits of the decompression engine.
This protection is only afforded when the hash Ugarit asks for is not
tampered with. Most hashes are obtained from within other blocks,
which are therefore safe unless that block has been tampered with; the
nature of the hash tree conveys the trust in the hashes up to the
root. The root hashes are stored in the vault as "tags", which an
vault modifier could alter at will. Therefore, the tags cannot be
trusted if somebody might modify the vault. This is why Ugarit
prints out the snapshot hash and the root directory hash after
performing a snapshot, so you can record them securely outside of the
vault.
The most likely threat posed by vault modifiers is that they could
simply corrupt or delete all of your vault, without needing to know
any encryption keys.
Recommendations: Secure your vaults against modifiers, by whatever
means possible. If vault modifiers are still a potential threat,
write down a log of your root directory hashes from each snapshot, and keep
it safe. When extracting your backups, use the ls -ll command in the
interface to check the "contents" hash of your snapshots, and check
they match the root directory hash you expect.
Process inspectors
These folks can attach debuggers or similar tools to running
processes, such as Ugarit itself.
Ugarit backend processes only see encrypted data, so people who can
attach to that process gain the powers of vault snoopers and
modifiers, and the same conditions apply.
People who can attach to the Ugarit process itself, however, will see
the original unencrypted content of your filesystem, and will have
full access to the encryption keys and hashing keys stored in your
Ugarit configuration. When Ugarit is running with sufficient
permissions to restore backups, they will be able to intercept and
modify the data as it comes out, and probably gain total write access
to your entire filesystem in the process.
Recommendations: Ensure that Ugarit does not run under the same user
ID as untrusted software. In many cases it will need to run as root in
order to gain unfettered access to read the filesystems it is backing
up, or to restore the ownership of files. However, when all the files
it backs up are world-readable, it could run as an untrusted user for
backups, and where file ownership is trivially reconstructible, it can
do restores as a limited user, too.
Attackers in the source filesystem
These folks create files that Ugarit will back up one day. By having
write access to your filesystem, they already have some level of
power, and standard Unix security practices such as storage quotas
should be used to control them. They may be people with logins on your
box, or more subtly, people who can cause servers to writes files;
somebody who sends an email to your mailserver will probably cause
that message to be written to queue files, as will people who can
upload files via any means.
Such attackers might use up your available storage by creating large
files. This creates a problem in the actual filesystem, but that
problem can be fixed by deleting the files. If those files get
stored into Ugarit, then they are a part of that snapshot. If you
are using a backend that supports deletion, then (when I implement
snapshot deletion in the user interface) you could delete that entire
snapshot to recover the wasted space, but that is a rather serious
operation.
More insidiously, such attackers might attempt to abuse a hash
collision in order to fool the vault. If they have a way of creating
a file that, for instance, has the same hash as your shadow password
file, then Ugarit will think that it already has that file when it
attempts to snapshot it, and store a reference to the existing
file. If that snapshot is restored, then they will receive a copy of
your shadow password file. Similarly, if they can predict a future
hash of your shadow password file, and create a shadow password file
of their own (perhaps one giving them a root account with a known
password) with that hash, they can then wait for the real shadow
password file to have that hash. If the system is later restored from
that snapshot, then their chosen content will appear in the shadow
password file. However, doing this requires a very fundamental break
of the hash function being used.
Recommendations: Think carefully about who has write access to your
filesystems, directly or indirectly via a network service that stores
received data to disk. Enforce quotas where appropriate, and consider
not backing up "queue directories" where untrusted content might
appear; migrate incoming content that passes acceptance tests to an
area that is backed up. If necessary, the queue might be backed up to
a non-snapshotting system, such as rsyncing to another server, so that
any excessive files that appear in there are removed from the backup
in due course, while still affording protection.
Acknowledgements
The Ugarit implementation contained herein is the work of Alaric
Snell-Pym and Christian Kellermann, with advice, ideas, encouragement
and guidance from many.
The original idea came from Venti, a content-addressed storage system
from Plan 9. Venti is usable directly by user applications, and is
also integrated with the Fossil filesystem to support snapshotting the
status of a Fossil filesystem. Fossil allows references to either be
to a block number on the Fossil partition or to a Venti key; so when a
filesystem has been snapshotted, all it now contains is a "root
directory" pointer into the Venti archive, and any files modified
therafter are copied-on-write into Fossil where they may be modified
until the next snapshot.
We're nowhere near that exciting yet, but using FUSE, we might be able
to do something similar, which might be fun. However, Venti inspired
me when I read about it years ago; it showed me how elegant
content-addressed storage is. Finding out that the Git version control
system used the same basic tricks really just confirmed this for me.
Also, I'd like to tip my hat to Duplicity. With the changing economics
of storage presented by services like Amazon S3 and rsync.net, I
looked to Duplicity as it provided both SFTP and S3 backends. However,
it worked in terms of full and incremental backups, a model that I
think made sense for magnetic tapes, but loses out to
content-addressed snapshots when you have random-access
media. Duplicity inspired me by its adoption of multiple backends, the
very backends I want to use, but I still hungered for a
content-addressed snapshot store.
I'd also like to tip my hat to Box Backup. I've only used it a little,
because it requires a special server to manage the storage (and I want
to get my backups *off* of my servers), but it also inspires me with
directions I'd like to take Ugarit. It's much more aware of real-time
access to random-access storage than Duplicity, and has a very
interesting continuous background incremental backup mode, moving away
from the tape-based paradigm of backups as something you do on a
special day of the week, like some kind of religious observance. I
hope the author Ben, who is a good friend of mine, won't mind me
plundering his source code for details on how to request real-time
notification of changes from the filesystem, and how to read and write
extended attributes!
Moving on from the world of backup, I'd like to thank the Chicken Team
for producing Chicken Scheme. Felix and the community at #chicken on
Freenode have particularly inspired me with their can-do attitudes to
combining programming-language elegance and pragmatic engineering -
two things many would think un-unitable enemies. Of course, they
didn't do it all themselves - R5RS Scheme and the SRFIs provided a
solid foundation to build on, and there's a cast of many more in the
Chicken community, working on other bits of Chicken or just egging
everyone on. And I can't not thank Henry Baker for writing the seminal
paper on the technique Chicken uses to implement full tail-calling
Scheme with cheap continuations on top of C; Henry already had my
admiration for his work on combining elegance and pragmatism in linear
logic. Why doesn't he return my calls? I even sent flowers.
A special thanks should go to Christian Kellermann for porting Ugarit
to use Chicken 4 modules, too, which was otherwise a big bottleneck to
development, as I was stuck on Chicken 3 for some time! And to Andy
Bennett for many insightful conversations about future directions.
Thanks to the early adopters who brought me useful feedback, too!
And I'd like to thank my wife for putting up with me spending several
evenings and weekends and holiday days working on this thing...
Version history
* 2.0: Archival mode [dae5e21ffc], and to support its integration
into Ugarit, implemented typed tags [08bf026f5a], displaying tag
types in the VFS [30054df0b6], refactoring the Ugarit internals
[5fa161239c], made the storage of logs in the vault better
[68bb75789f], made it possible to view logs from within the VFS
[4e3673e0fe], supported hidden tags [cf5ef4691c], recording
configuration information in the vault (and providing instant
notification if your vault hashing/encryption setup is incorrect,
thanks to a clever idea by Andy Bennett) [0500d282fc], rearranged
how local caching is handled [b5911d321a], and added support for
the history of a snapshot or archive tag to have arbitrary
branches and merges [a987e28fef], which (as a side-effect)
improved the performance of running "ls" in long snapshot
histories [fcf8bc942a]. Also added an sqlite backend
[8719dfb84f], which makes testing easier but is useful in its own
right as it's fully-featured and crash-safe, while storing the
vault in a single file; and improved the appearance of the
explore mode ls command, as the VFS layout has become more
complex with the new log/properties views and all the archive
mode stuff.
* 1.0.9: More humane display of sizes in explore's directory
listings, using low-level I/O to reduce CPU usage. Myriad small
bug fixes and some internal structural improvements.
* 1.0.8: Bug fixes to work with the latest chicken master, and
increased unit test coverage to test stuff that wasn't working
due to chicken bugs. Looking good!
* 1.0.7: Fixed bug with directory rules (errors arose when files
were skipped). I need to improve the test suite coverage of
high-level components to stop this happening!
* 1.0.6: Fixed missing features from v1.0.5 due to a fluffed merge
(whoops), added tracking of directory sizes (files+bytes) in the
vault on snapshot and the use of this information to display
overall percentage completion when extracting. Directory sizes
can be seen in the explore interface when doing "ls -l" or "ls -ll".
* 1.0.5: Changed the VFS layout slightly, making the existence of
snapshot objects explicit (when you go into a tag, then go into a
snapshot, you now need to go into "contents" to see the actual
file tree; the snapshot object itself now exists as a node in the
tree). Added traverse-vault-* functions to the core API, and tests
for same, and used traverse-vault-node to drive the cd and get
functions in the interactive explore mode (speeding them up in the
process!). Added "extract" command. Added a progress reporting
callback facility for snapshots and extractions, and used it to
provide progress reporting in the front-end, every 60 seconds or
so by default, not at all with -q, and every time something
happens with -v. Added tab completion in explore mode.
* 1.0.4: Resurrected support for compression and encryption and SHA2
hashes, which had been broken by the failure of the
autoload egg to continue to work as it used to. Tidying
up error and ^C handling somewhat.
* 1.0.3: Installed sqlite busy handlers to retry when the database is
locked due to concurrent access (affects backend-fs, backend-cache,
and the file cache), and gained an EXCLUSIVE lock when locking a
tag in backend-fs; I'm not clear if it's necessary, but it can't
hurt.
BUGFIX: Logging of messages from storage backends wasn't
happening correctly in the Ugarit core, leading to errors when the
cache backend (which logs an info message at close time) was closed
and the log message had nowhere to go.
* 1.0.2: Made the file cache also commit periodically, rather than on
every write, in order to improve performance. Counting blocks and
bytes uploaded / reused, and file cache bytes as well as hits;
reporting same in snapshot UI and logging same to snapshot
metadata. Switched to the posix-extras egg and ditched our own
posixextras.scm wrappers. Used the parley egg in the ugarit
explore CLI for line editing. Added logging infrastructure,
recording of snapshot logs in the snapshot. Added recovery from
extraction errors. Listed lock state of tags in explore
mode. Backend protocol v2 introduced (retaining v1 for
compatability) allowing for an error on backend startup, and logging
nonfatal errors, warnings, and info on startup and all protocol
calls. Added ugarit-archive-admin command line interface to
backend-specific administrative interfaces. Configuration of the
splitlog backend (write protection, adjusting block size and logfile
size limit and commit interval) is now possible via the admin
interface. The admin interface also permits rebuilding the metadata
index of a splitlog vault with the reindex! admin command.
BUGFIX: Made file cache check the file hashes it finds in the
cache actually exist in the vault, to protect against the case
where a crash of some kind has caused unflushed changes to be
lost; the file cache may well have committed changes that the
backend hasn't, leading to references to nonexistant blocks. Note
that we assume that vaults are sequentially safe, eg if the
final indirect block of a large file made it, all the partial
blocks must have made it too.
BUGFIX: Added an explicit flush! command to the backend
protocol, and put explicit flushes at critical points in higher
layers (backend-cache, the vault abstraction in the Ugarit
core, and when tagging a snapshot) so that we ensure the blocks we
point at are flushed before committing references to them in the
backend-cache or file caches, or into tags, to ensure crash
safety.
BUGFIX: Made the splitlog backend never exceed the file size limit
(except when passed blocks that, plus a header, are larger than
it), rather than letting a partial block hang over the 'end'.
BUGFIX: Fixed tag locking, which was broken all over the
place. Concurrent snapshots to the same tag should now block for
one another, although why you'd want to *do* that is questionable.
BUGFIX: Fixed generation of non-keyed hashes, which was
incorrectly appending the type to the hash without an outer
hash. This breaks backwards compatability, but nobody was using
the old algorithm, right? I'll introduce it as an option if
required.
* 1.0.1: Consistency check on read blocks by default. Removed warning
about deletions from backend-cache; we need a new mechanism to
report warnings from backends to the user. Made backend-cache and
backend-fs/splitlog commit periodically rather than after every
insert, which should speed up snapshotting a lot, and reused the
prepared statements rather than re-preparing them all the
time.
BUGFIX: splitlog backend now creates log files with
"rw-------" rather than "rwx------" permissions; and all sqlite
databases (splitlog metadata, cache file, and file-cache file) are
created with "rw-------" rather then "rw-r--r--".
* 1.0: Migrated from gdbm to sqlite for metadata storage, removing the
GPL taint. Unit test suite. backend-cache made into a separate
backend binary. Removed backend-log.
BUGFIX: file caching uses mtime *and*
size now, rather than just mtime. Error handling so we skip objects
that we cannot do something with, and proceed to try the rest of the
operation.
* 0.8: decoupling backends from the core and into separate binaries,
accessed via standard input and output, so they can be run over SSH
tunnels and other such magic.
* 0.7: file cache support, sorting of directories so they're archived
in canonical order, autoloading of hash/encryption/compression
modules so they're not required dependencies any more.
* 0.6: .ugarit support.
* 0.5: Keyed hashing so attackers can't tell what blocks you have,
markers in logs so the index can be reconstructed, sha2 support, and
passphrase support.
* 0.4: AES encryption.
* 0.3: Added splitlog backend, and fixed a .meta file typo.
* 0.2: Initial public release.
* 0.1: Internal development release.