support new ENV vars to get secret values through a file or a command
We want to allow passing a secret not only directly through the
environment value, but also indirectly through a file path, an open
file descriptor or a command that can write it to standard out.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
move remaining client tools to pbs-tools/datastore
pbs-datastore now ended up depending on tokio after all, but
that's fine for now
for the fuse code I added pbs-fuse-loop (has the old
fuse_loop and its 'loopdev' module)
ultimately only binaries should depend on this to avoid the
library link
the only thins remaining to move out the client binary are
the api method return types, those will need to be moved to
pbs-api-types...
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Thomas Lamprecht [Tue, 20 Jul 2021 17:58:03 +0000 (19:58 +0200)]
buildsys: run test before compile to avoid clobbering the openid build binaries
dh_auto_test also checks for the build flags used, including any
`--cfg`, so it rebuilds and overwrites our carefully assembled daemon
binaries with openid support as it is run after build and before
install.
So manually ensure the order of first test then build (argh, hackes
of hackes >.<)
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Thomas Lamprecht [Mon, 19 Jul 2021 14:32:48 +0000 (16:32 +0200)]
issue banner: avoid depending on proxmox crate for hostname
While this slightly duplicates code we just do not profit from the
central, lazy static variant here, as that is only really useful in
daemons to avoid doing frequent syscalls there.
proxmox just pull in far to much (e.g., tokio) and duplicating that
one line of simple code has no real maintenance cost, so just go for
that and use the nix crate directly.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
since the api call always starts a real worker, we cannot have a
preview. It would also be very hard to show that for all groups in a
non-confusing way. We reuse the pbsPruneInputPanel and add the dry-run
field there conditionally.
api: admin/datastore: add new 'prune-datastore' api call
to prune the whole datastore at once, with the given parameters.
We need a new api call since this can take a while and we need to start
a worker for this. The exisiting api call returns a list of removed/kept
snapshots and is synchronous.
server/prune_job: add 'keep_all' logic to 'prune_datastore'
it is the same as when pruning single groups.
for prune_jobs, we never start the worker if there is no prune option set.
but if we want to call 'prune_datastore' from somewhere else, we
have to check it here again
fix #3526: correctly filter tasks with 'since' and 'until'
The previous assumption was that the Tasks returned by the Iterator are
sorted by the starttime, but that is not actually the case, and
could never have been, since we append the tasks into the log when
they are finished (not started) and running tasks are always iterated
first.
To correctly filter (and simplify the the api call) we forgo the
combinators, and use a for loop instead. This way we only have to do
the since/until checks only once per Task, but have to do the
start/limit counting ourselves.
Stefan Reiter [Tue, 13 Jul 2021 09:23:21 +0000 (11:23 +0200)]
fix #3515: file-restore-daemon: allow LVs/PVs with dash in name
LVM replaces any dashes '-' in an LV or PV name with two '--' for the
created device node in /dev/mapper/ to distinguish the seperating
character between the PV and LV name.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Stefan Reiter [Tue, 13 Jul 2021 09:23:20 +0000 (11:23 +0200)]
file-restore: increase lock timeout on QEMU map
This lock is held during VM startup, so that multiple calls will not
start VMs twice. But this means that the timeout needs to incorporate
the time it might take a VM to boot, so increase it quite a bit.
This could previously lead to "interrupted system call" errors when
accessing backups with many disks.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
ui: tape/BackupOverview: increase timeout for media-set content
a single catalog can be over 100MiB, and a media-set can have multiple
catalogs to read (no technical upper limit). On slow disks, this can
take much longer than 30 seconds (the default timeout).
The real solution would be to have some kind of index only for the gui
relevant part, e.g. a table in the beginning of the catalog, or
alternatively a seperate file with that info. Until we have such a
solution increase the timeout as a stopgap.
Stefan Reiter [Mon, 12 Jul 2021 16:30:47 +0000 (18:30 +0200)]
auth: 'crypt' is not thread safe
According to crypt(3):
"crypt places its result in a static storage area, which will be
overwritten by subsequent calls to crypt. It is not safe to call crypt
from multiple threads simultaneously."
This means that multiple login calls as a PBS-realm user can collide and
produce intermittent authentication failures. A visible case is for
file-restore, where VMs with many disks lead to just as many auth-calls
at the same time, as the GUI tries to expand each tree element on load.
Instead, use the thread-safe variant 'crypt_r', which places the result
into a pre-allocated buffer of type 'crypt_data'. The C struct is laid
out according to 'lib/crypt.h.in' and the man page mentioned above.
Use the opportunity and make both arguments to the rust 'crypt' function
take a &[u8].
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
tape: changer: sg_pt: query element types separately
Some changers do not like the DVCID bit when querying non-drives,
this includes when querying 'all' elements.
To circumvent this, we query each type by itself (like mtx does it),
and only add the DVCID bit for drives (Data Transfer Elements).
Reported by a user in the forum:
https://forum.proxmox.com/threads/ibm-3584-ts3500-support.92291/
and limit to 1000 elements per request.
(Because some changers limit that request with the options we set)
instead of checking if the data len was equal to the allocation_len
for getting more data, we count the returned elements and compare
that with the number we requested
New kernel has stricter checks on tmpfs with stick-bit on directories, so some
commands (i.e. proxmox-tape changer status) fails when executed as root, because
permission checks fails when locking the drive.
This patch move the drive locks to /run/proxmox-backup/drive-lock.
Note: This is incompatible to old locking mechmanism, so users may not
run tape backups during update (or running backup can fail).
Thomas Lamprecht [Mon, 12 Jul 2021 11:15:17 +0000 (13:15 +0200)]
buildsys: try to avoid duplicate build due to "phony" docs dependency
Make docs target depend directly on the some docs-only required
binaries and add a new intermediate ".do-cargo-build" target that is
explicitly not a PHONY target.
That avoids one extra set of full builds.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Thomas Lamprecht [Mon, 12 Jul 2021 05:57:37 +0000 (07:57 +0200)]
ui: datastore content: style edit notes pencil like action-col icon
as those have a hover effect and use dark-grey vs. the quite "harsh"
looking plain black. We need to override the margin though, as else
the floated layout adds another line.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Thomas Lamprecht [Mon, 12 Jul 2021 05:54:41 +0000 (07:54 +0200)]
ui: update group notes: fix obj access and rewrite to async
eslint is configured to not allow using quoted object keys if they
could be just passed in dot notation, e.g.,
wrong: `group["comment"]`
good: `group.comment`
It's not a big problem but eslint fails the build with the wrong one,
so this needs to be fixed anyway..
Also, rewrite to async, shorter and less indentation
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Stefan Reiter [Thu, 8 Jul 2021 14:45:28 +0000 (16:45 +0200)]
ui: add support for notes on backup groups
Currently done a little bit hacky in a seperate API call following the
initial list_snapshots, as we previously didn't call list_groups at all
and instead calculated the groups from the snapshots.
This calls it async and updates the view with group comments when data
arrives. The editor is simply reused with the 'group-notes' API call,
since the semantics are the same.
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Stefan Reiter [Thu, 8 Jul 2021 14:45:27 +0000 (16:45 +0200)]
api: add support for notes on backup groups
Stored in atomically-updated 'notes' file in backup group directory.
Available via dedicated GET/PUT API calls, as well as the first line
being included in list_groups (similar to list_snapshots).
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com> Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>