The Python app that "hooks" into Proxmox.
This is a clone of vzdump-hook-script.pl that is written in Python, and that offers a lot more configurability.
Different phases of the vzdump
backup can be hooked into, and things can be run.
The app also logs script output in realtime -- useful when using a long-running process (like rclone
for example),
and you want to see progressive timestamping against its output.
⚠️ NOTE! Version 2.0.0 introduces a breaking change to the config file format! See below.
Having hooks into each part of vzdump
backups is very useful, especially to detect failures (or success) of the
different backup phases.
For example, it can be used in concert with Healthchecks.io, Mailrise/Apprise, or other apps, to receive status notifications:
production:
job-start:
mode: script
run:
- "curl -fsS -m 10 --retry 5 -o /dev/null https://your.healthchecks.server/ping/xxx/vzdump-backups/start"
job-abort:
mode: script
run:
- "curl -fsS -m 10 --retry 5 -o /dev/null https://your.healthchecks.server/ping/xxx/vzdump-backups/fail"
backup-abort:
mode: script
run:
- "curl -fsS -m 10 --retry 5 -o /dev/null https://your.healthchecks.server/ping/xxx/vzdump-backups/fail"
Maybe you'd like to offsite-sync your backups on job completion:
job-end:
mode: script
run:
- "ssh some.host rclone sync --checkers 32 --transfers 16 --dscp cs1 --stats-log-level NOTICE --stats-unit=bits --stats=2m /mnt/pbs-backups remote.host:pbs-rsync"
- "curl -fsS -m 10 --retry 5 -o /dev/null https://your.healthchecks.server/ping/xxx/vzdump-backups"
Or maybe as @lloydbayley uses it, via curl
:
It actually changes the RGB lights in my rack when a backup happens so it's purely decorative but I like to automate things.
Anything that can be run on the CLI, you can use here.
Instead of using mode: script
, you can use mode: shell
, and anything configured will be run through a shell. This is
directly equivalent to the shell
argument to Python's subprocess.Popen
.
What's the benefit here? To quote the Python documentation:
If shell is True, the specified command will be executed through the shell. This can be useful if you are using Python primarily for the enhanced control flow it offers over most system shells and still want convenient access to other shell features such as shell pipes, filename wildcards, environment variable expansion, and expansion of ~ to a user’s home directory.
The shell used is /bin/sh
.
⚠️ WARNING! Read Python's Security Considerations section before usingshell
!
proxmox-grapple
will expose the vzdump
parameters to the following environment variables:
GRAPPLE_PHASE
GRAPPLE_MODE
GRAPPLE_VMID
And all arguments in one:
GRAPPLE_ALL_ARGS
These can then be used in your config as you see fit. Note you'll need to use shell
mode, or use binaries that can
read/expand/interpolate environment variables for you.
An example:
GRAPPLE_PHASE=backup-start
GRAPPLE_MODE=stop
GRAPPLE_VMID=102
GRAPPLE_ALL_ARGS=backup-start stop 102
The recommended way to install proxmox-grapple
is to use pipx.
After getting pipx
installed, simply run:
username@proxmox:~$ pipx install proxmox-grapple
Please don't use pip system-wide.
You can of course also install it using classic virtualenvs.
Once installed, you need to tell Proxmox (specifically vzdump
) to use the app. Edit /etc/vzdump.conf
and add or edit
the script
setting:
script: /home/username/.local/bin/proxmox-grapple
No other changes to the file is needed.
proxmox-grapple
is configured with a YAML-style file. An example:
production:
job-end:
mode: script
run:
- echo 'hi'
- sleep 1
- echo 'there'
- echo 'This is a test.'
backup-end:
mode: script
run:
extract:
enabled: false
source_directory: /tmp
destination_directory: /tmp
# exclude_storeids:
job-abort:
mode: shell
run:
- echo 'your strange command' | tee some_logfile.txt
Each top-level key should match the different vzdump
phases. The currently recognised phases are:
job-init
job-start
job-end
job-abort
backup-start
backup-end
backup-abort
log-end
pre-stop
pre-restart
post-restart
All of these are optional, and if not configured, are ignored.
If they are configured, the accompanying sub-keys are required:
mode
run
⚠️ The extract argument is currently not tested, and should be treated as a proof-of-concept only.
The default location for the configuration is /etc/proxmox_grapple.yml
, or proxmox_grapple.yml
in the current
working directory, but this can also be specified on the commandline.
If a non-absolute path is given, Dynaconf will iterate upwards: it will look at each parent up to the root of the
system. For each visited folder, it will also try looking inside a /config
folder.
There is a mode, --dump-config
, that reads all possible configuration and then prints it out for validation purposes.
It can also use different configuration settings based on arbitrary environment names (eg. production
, lab
, etc.) It
uses the Dynaconf Python project for configuration, so using environment variables to
choose the environment and any other configuration option is possible.
⚠️ The default environment isproduction
For example, to choose a different configuration environment, set the environment variable ENV_FOR_DYNACONF=lab
:
root@proxmox:~# ENV_FOR_DYNACONF=lab proxmox-grapple --dump-config
To enable the use of the two run modes, I've decided to change the schema of the config file that needs to be updated.
production:
backup-start:
script:
- echo hi
production:
backup-start:
mode: script
run:
- echo hi
proxmox-grapple
supports the following VE versions:
VE version | Debian version | Python version | VE EoL |
---|---|---|---|
8 | 12 (Bookworm) | 3.11 | TBA |
7 | 11 (Bullseye) | 3.9 | 2024-07 |