in case the garbage_collection errors out, we never set the in-memory
state, so if it failed, the last 'good' starttime was considered
for the schedule
this could lead to the job running every minute instead of the
correct schedule
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
instead of manually, this has the advantage that we now set
the jobstate correctly and can return with an error if it is
currently running (instead of failing in the task)
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
index files that were smaller than their respective header size,
would fail with
"failed to fill whole buffer"
instead now check explicitely for the size and fail with
"index too small (size)"
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
try do reduce some unecessary lines, make match arms more precise so
one can faster see what's actually happening.
Also, avoid
> return Err(format_err!(...))
stuff, just use bail!()
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
re-use the future we already have for task log rotation to trigger
it.
Move the FileLogger in ApiConfig into an Arc, so that we can actually
update it and REST using the new one.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
so that we can easily get the main PID of the last recently launched
daemon. Will be used to get the control socket of that one for access
lgo rotate in a future patch
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
this is internal for now, use the comanndo socket struct
implementation, and ideally not a new one but the existing ones
created in the proxy and api daemons.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Allows to extend the use of that socket in the future, e.g., for log
rotate re-open signaling.
To reflect this we use a more general name, and change the commandos
to a more clear namespace.
Both are actually somewhat a breaking change, but the single real
world issue it should be able to cause is, that one won't be able to
stop task from older daemons, which still use the older abstract
socket name format.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
This is a preparatory step to replace the task control socket with it
and provide a "reopen log file" command for the rest server.
Kept it simple by disallowing to register new commands after the
socket gets spawned, this avoids the need for locking.
If we really need that we can always wrap it in a Arc<RWLock<..>> or
something like that, or even nicer, register at compile time.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
writing to a file can explode quite easily.
time formatting to rfc3339 should be more robust, but it has a few
conditions where it could fail, so catch that too (and only really
do it if required).
The writes to stdout are left as is, it normally is redirected to
journal which is in memory, and thus breaks later than most stuff,
and at that point we probably do not care anymore anyway.
It could make sense to actually return a result here..
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
We renamed the last one always to a file without compression
extension, even if it was .zst previously. So always add the correct
ending to the new last one, if compress was true.
Further, we cannot detect if there'd be a compression required if we
rotated (renamed) it already to the file with .zst included.
So check on rotation itself if it would be a "no .zst" -> ",zst"
transition, and call compress there.
it really should be OK now *knocking wood*
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
by requiring
- Datastore.Backup permission for target datastore
- Remote.Read permission for source remote/datastore
- Datastore.Prune if vanished snapshots should be removed
- Datastore.Modify if another user should own the freshly synced
snapshots
reading a sync job entry only requires knowing about both the source
remote and the target datastore.
note that this does not affect the Authid used to authenticate with the
remote, which of course also needs permissions to access the source
datastore.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
instead of hard-coding 'backup@pam'. this allows a bit more flexibility
(e.g., syncing to a datastore that can directly be used as restore
source) without overly complicating things.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
no idea why I added it as "delete", for all other such operations we
use the "remove" sub-command...
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
again, base idea copied off PVE, but, we safe the information about
which pending version we send a mail out already in a separate
object, to keep the api return type APTUpdateInfo clean.
This also makes a few things a bit easier, as we can update the
package status without saving/restoring the notify information.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
apt changes some of its state/cache also if it errors out, most of
the time, so we actually want to print both, stderr and stdout.
Further, only warn if its exit code is non-zero, for the same
rationale, it may bring updates available even if it errors (e.g.,
because a future pbs-enterprise repo is additionally configured but
not accessible).
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
it's not used anywhere, and not needed either until the day we might
implement push syncs.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
equivalent to verifying a whole datastore, except for reading job
(entries), which is accessible to regular Datastore.Audit/Backup users
as well.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
for verifying a whole datastore. Datastore.Backup now allows verifying
only backups owned by the triggering user.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
they are returned when reading the manifest, which just requires
Datastore.Audit as well. Datastore.Read is for reading backup contents,
not metadata.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
even for otherwise unprivileged users.
since effective privileges of an API token are always intersected with
those of their owning user, this does not allow an unprivileged user to
elevate their privileges in practice, but avoids the need to involve a
privileged user to deploy API tokens.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
a user should be allowed to read/list/overwrite backups owned by their
own tokens, but a token should not be able to read/list/overwrite
backups owned by their owning user.
when changing ownership of a backup group, a user should be able to
transfer ownership to/from their own tokens if the backup is owned by
them (or one of their tokens).
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
since it's not possible to extend existing structs, UserWithTokens
duplicates most of user::User.. to avoid duplicating user::ApiToken as
well, this returns full API token IDs, not just the token name part.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
in most generic places. this is accompanied by a change in
RpcEnvironment to purposefully break existing call sites.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
with an optional Tokenname, appended with '!' as delimiter in the string
representation like for PVE.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Make it more clear that removed files are chunks (not indexes or
something like that, user cannot know that we do not touch them here)
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
fixes commit b4fb262335, which copied
over the "Removed bad files:" block, but only adapted the log text,
not the actual variable.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
instead of prerotating 1000 tasks
(which resulted in 2 writes each time an active worker was finished)
simply append finished tasks to the archive (which will be rotated)
page cache should be good enough so that we can get the task logs fast
since existing installations might have an 'index' file, we
still have to read tasks from there, but only if it exists
this simplifies the TaskListInfoIterator a good amount
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
by moving the properties of the storage status out again to the top
level object
also introduce proper structs for the types used, to get type-safety
and better documentation for the api calls
this changes the backup counts from an array of [groups,snapshots] to
an object/struct with { groups, snapshots } and include 'other' types
(though we do not have any at this moment)
this way it is better documented
this also adapts the ui code to cope with the api changes
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
saves files mtime as i64 instead of u64 which enables backup of
files with negative mtime
the catalog_decode_i64 is compatible to encoded u64 values (if < 2^63)
but not reverse, so all "old" catalogs can be read with the new
decoder, but catalogs that contain negative mtimes will decode wrongly
on older clients
also remove the arbitrary maximum value of 2^63 - 1 for
encode_u64 (we just use up to 10 bytes now) and correctly
decode them and update the comments accordingly
adds also test for i64 encode/decode and for compatibility between
u64 encode and i64 decode
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
fixes commit 16f9f244cf which extended
the return schema of the status API but did not adapted the client
status command to that.
Simply define our own tiny return schema and use that.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
we never actually compressed any files, since we only looked at
the extension:
* if it was 'zst' (which was always true for newly rotated files), we
would not compress it
* even if it was not 'zst', we compressed it inplace, never adding '.zst'
(possibly compressing them multiple times as zstd)
now we add new rotated files simply as '.X' and add a 'target' to the
compress fn, where we rename it to (but now we have to unlink the source
path)
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
commit a4915dfc2b made a wrong fix, as
it did not observed that the last expressions was done under the
invariant that we had a last verification result, because if none
could be loaded we already returned true (include).
It thus broke the case for "never re-verify", which is important when
using multiple schedules, a more high frequent one for new,
unverified snapshots, and a low frequency to re-verify older snapshots,
e.g., monthly.
Fix this case again, rework the code to avoid this easy to oversee
invariant. Use a nested match to better express the implication of
each setting, and add some comments.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
and load it again when opening it
this way we can persist the status of the last garbage collect across
daemon reloads and reboots
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
This is intended for when the server needs to do requests on
arbitrary, non PBS, external HTTP resources.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
we're a bit strict here what we accept, rather than changing that
lets do it like PVE/PMG and uppercase.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
This is indented to be used for the PVE storage library, replacing
the missuse of the much more expensive status API call.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
For proxmox packages it works the same way as PVE, by retrieving the
changelog URL and issuing a HTTP GET to it, forwarding the output to the
client. As this is only supposed to be a workaround removed in the
future, a simple block_on is used to avoid async.
For debian packages we can simply call 'apt-get changelog' and forward
it's output.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
As always, libapt is mocking us with complexity, but we can get the
approximate result we want by retrieving dependencies of all
to-be-updated packages and then seeing if they are missing.
If they are, we assume they will be installed.
For this, query_detailed_info is extended to allow reading details for
non-installed packages, and this is also exposed in
list_installed_apt_packages via 'all_versions_for'. This is necessary so
we can retrieve changelogs for such packages.
Note that we cannot retrieve all that information all the time, as
querying details for packages that aren't installed takes a rather long
time.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Avoids custom hardcoded logic, but can only be used for debian packages
as of now. Adds a FIXME to switch over to use --print-uris only once our
package repos support that changelog format.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
To get package details for a specific version instead of only the
candidate.
Also cleanup filter function with extra struct instead of unnamed &str
parameters.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
...to avoid having the tools:: module depend on api2.
The get_string function is based directly on hyper and thus relatively
simple, not supporting redirects for example.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
gets rid of the return value and moving around of the zip
and decoder data
avoids cloning the path prefix on every recursion
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
by using the new ZipEncoder and recursively add files to it
the zip only contains directories, normal files and hardlinks (by simply
copying the content), no symlinks, etc.
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
similar to StdChannelWriter, but implements AsyncWrite and sends
to a tokio::sync::mpsc::Sender
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
This modules contains the 'ZipEncoder' struct, which wraps an async writer,
to create a ZIP archive on the fly
To create a ZIP file, have a target that implements AsyncWrite,
give it to ZipEncoder::new, add entries via 'add_entry' and
at the end, call 'finish'
for now, this does not implement compression (uses ZIPs STORE mode), and
does not support empty directories or hardlinks (or any other special
files)
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Fixes a bug in which the userid of the ticket cache is updated,
when a user connects, but the ticket itself is not.
This means a newly connected user has a previously connected
user's ticket and thus, cannot do anything, as the client will
attempt to use the invalid ticket.
e.g. if john@pbs connected to the server first, followed by
mike@pbs, the following would be stored in the ticket cache.
{
"localhost": {
"mike@pbs": {
"ticket": "PBS:john@pbs:AAAA",
"timestamp": 1601039326,
"token": "BBBB"
}
}
}
Signed-off-by: Dylan Whyte <d.whyte@proxmox.com>
The first rotation is normally the one still opened by one or more
processes for writing, so it must NOT be replaced, removed, ..., as
this then makes the remaining logging, until those processes are
noticed that they should reopen the logfile due to rotation, goes
into nirvana, which is far from ideal for a log.
Only rotating (renaming) is OK for this active file, as this does not
invalidates the file and keeps open FDs intact.
So start compressing with the second rotation, which should be clear
to use, as all writers must have been told to reopen the log during
the last rotation, reopen is a fast operation and normally triggered
at least day ago (at least if one did not dropped the state file
manually), so we are fine to archive that one for real.
If we plan to allow faster rotation the whole rotation+reopen should
be locked, so that we can guarantee that all writers switched over,
but this is unlikely to be needed.
Again, this is was logrotate sanely does by default since forever.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
this is not the job of logrotate, and the real 20+ years battle
tested log rotate binary does not do so either as it's actually
pretty dangerous.
If we "replace" the file we break any logger which already opened a
new one here, e.g., a dameon starting up, and thus that writer would
log to nirvana.
It's the job of a logger to create a file if not existing, it makes
no sense to do it here.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
To cater to the paranoid, a new datastore-wide setting "verify-new" is
introduced. When set, a verify job will be spawned right after a new
backup is added to the store (only verifying the added snapshot).
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Force consumers to use the lookup_datastore method instead of
potentially opening a datastore twice, and pass the config we have
already loaded into open_with_path, removing the need for open(1).
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Commit 9070d11f4c introduced this change for other call sites,
assuming it is correct, this one was missed.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
and use that in ApiConfig to avoid that it is owned by root if the
proxmox-backup-api process creates it first.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
for now log auth errors also to the syslog, on a protected (LAN
and/or firewalled) setup this should normally happen due to
missconfiguration, not tries to break in.
This reduces syslog noise *a lot*. A current full journal output from
the current boot here has 72066 lines, of which 71444 (>99% !!) are
"successful auth for user ..." messages
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
reuse the FileLogger module in append mode.
As it implements write, which is not thread safe (mutable self) and
we use it in a async context we need to serialize access using a
mutex.
Try to use the same format we do in pveproxy, namely the one which is
also used in apache or nginx by default.
Use the response extensions to pass up the userid, if we extract it
from a ticket.
The privileged and unprivileged dameons log both to the same file, to
have a unified view, and avoiding the need to handle more log files.
We avoid extra intra-process locking by reusing the fact that a write
smaller than PIPE_BUF (4k on linux) is atomic for files opened with
the 'O_APPEND' flag. For now the logged request path is not yet
guaranteed to be smaller than that, this will be improved in a future
patch.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Add a generous limit now and return the correct error (414 URI Too
Long). Otherwise we could to pretty larger GET requests, 64 KiB and
possible bigger (at 64 KiB my simple curl test failed due to
shell/curl limitations).
For now allow a 3072 characters as combined length of URI path and
query.
This is conform with the HTTP/1.1 RFCs (e.g., RFC 7231, 6.5.12 and
RFC 2616, 3.2.1) which do not specify any limits, upper or lower, but
require that all server accessible resources mus be reachable without
getting 414, which is normally fulfilled as we have various length
limits for stuff which could be in an URI, in place, e.g.:
* user id: max. 64 chars
* datastore: max. 32 chars
The only known problematic API endpoint is the catalog one, used in
the GUI's pxar file browser:
GET /api2/json/admin/datastore/<id>/catalog?..&filepath=<path>
The <path> is the encoded archive path, and can be arbitrary long.
But, this is a flawed design, as even without this new limit one can
easily generate archives which cannot be browsed anymore, as hyper
only accepts requests with max. 64 KiB in the URI.
So rather, we should move that to a GET-as-POST call, which has no
such limitations (and would not need to base32 encode the path).
Note: This change was inspired by adding a request access log, which
profits from such limits as we can then rely on certain atomicity
guarantees when writing requests to the log.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
needs new proxmox dependency to get the RpcEnvironment changes,
adding client_ip getter and setter.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Rewrite most of the documentation to be more readable and correct
(according to the current implementations).
Add a table visualizing all different locks used to synchronize
concurrent operations.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Contains a link to the 'backup' module's doc, as that explains a lot
about the inner workings of PBS and probably marks a good entry point
for new readers.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Avoid races when updating manifest data by flocking a lock file.
update_manifest is used to ensure updates always happen with the lock
held.
Snapshot deletion also acquires the lock, so it cannot interfere with an
outstanding manifest write.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
The 'Ok::<_, Self::Error>(res)' type annotation was from a time where
we could not use async, and had a combinator here which needed
explicity type information. We switched over to async in commit
91e4587343 and, as the type annotation
is already included in the Future type, we can safely drop it.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Given the .pxarexclude file
foo
/bar
The following happens:
exclude: /foo
exclude: /bar
exclude: /subdir/foo
include: /subdir/bar
since the `/bar` line is an absolute path
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Removing a snapshot has some more safety checks which we don't want to
ignore when removing an entire group (i.e. locking the manifest and
notifying GC).
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
There's no point in having that as a seperate method, just parse the
thing into a struct and write it back out correctly.
Also makes further changes to the method simpler.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
If we can't acquire a lock (either because the snapshot disappeared, it
is about to be forgotten/pruned, or it is currently still running) skip
the snapshot. Hold the lock during verification, so that it cannot be
deleted while we are still verifying.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
...to avoid it being forgotten or pruned while in use.
Update lock error message for deletions to be consistent.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
To allow other reading operations on the base snapshot as well. No
semantic changes with this patch alone, as all other locks on snapshots
are exclusive.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
A removal can fail if the snapshot is already gone (this is fine, our
job is done either way) or we couldn't get a lock (also fine, it can't
be removed then, just warn the user so he knows what happened and why it
wasn't removed) - keep going either way.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
A snapshot that's currently being read can still appear in the prune
list, but should not be removed.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
This adds a change-owner command to proxmox-backup-client,
that allows a caller with datastore modify privileges
to change the owner of a backup-group.
Signed-off-by: Dylan Whyte <d.whyte@proxmox.com>
To untangle the server code from the actual backup
implementation.
It would be ideal if the whole backup/ dir could become its
own crate with minimal dependencies, certainly without
depending on the actual api server. That would then also be
used more easily to create forensic tools for all the data
file types we have in the backup repositories.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Used to not require access to the WorkerTask struct outside
the `server` and `api2` module, so it'll be easier to
separate those backup/server/client parts into separate
crates.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
This is only acquired in those two methods, both as shared. So it has
no use.
It seems, that it was planned in the past that the index deletion
should take the exclusive, while read and write takes the shared
flock on the index, as one can guess from the lock comments in commit
0465218953
But then later, in commit c8ec450e37)
the documented semantics where changed to use a temp file and do an
atomic rename instead for atomicity.
The reader shared flock on the index file was done inbetween,
probably as preparatory step, but was not removed again when strategy
was changed to using the file rename instead.
Do so now, to avoid confusion of readers and a useless flock.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
via HTTP2/backup reader protocol. they already could do so via the plain
HTTP download-file/.. API calls that the GUI uses, but the reader
environment required READ permission on the whole datastore instead of
just BACKUP on the backup group itself.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
a reader connection should not be allowed to read arbitrary chunks in
the datastore, but only those that were previously registered by opening
the corresponding index files.
this mechanism is needed to allow unprivileged users (that don't have
full READ permissions on the whole datastore) access to their own
backups via a reader environment.
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Fabian Grünbichler
Thomas Lamprecht
Dominik Csapak
Dietmar Maurer
Wolfgang Bumiller
Hannes Laimer
Stefan Reiter
Dylan Whyte