[pve-devel] applied: [PATCH container v2 2/6] api: status: move config locking from API handler into worker

[Thomas Lamprecht]

Am 30/01/2024 um 18:10 schrieb Friedrich Weber:
> Previously, container start/stop/shutdown/suspend would try to acquire
> the config lock in the API handler prior to forking a worker. If the
> lock was currently held elsewhere, this would block the API handler
> and thus the pvedaemon worker thread until the 10s timeout expired (or
> the lock could be acquired).
> 
> To avoid blocking the API handler, immediately fork off a worker
> process and try to acquire the config lock in that worker.
> 
> Patch best viewed with `git show -w`.
> 
> Suggested-by: Wolfgang Bumiller <w.bumiller at proxmox.com>
> Signed-off-by: Friedrich Weber <f.weber at proxmox.com>
> ---
> 
> Notes:
>     The diff is somewhat messy without `-w` -- couldn't come up with a
>     better way.
>     
>     new in v2
> 
>  src/PVE/API2/LXC/Status.pm | 91 ++++++++++++++++++--------------------
>  1 file changed, 42 insertions(+), 49 deletions(-)
> 
>

applied this one already, thanks!

btw. in general getting some early timeout sent as response to the initial
request would be good, but blocking the daemon worker is naturally a no-go.
And bringing in Anyevent somehow to make it async-like might be a PITA, so
for now this change is OK.

Oh, and it might be worth mentioning explicitly in the next release notes,
as it's a change in behavior that could theoretically throw up some
tooling that depends on the $action not failing due to locking if the
adapted endpoints returned – albeit IMO a bit silly to assume that (as the
actions can fail in other ways inside the worker anyway).