[pbs-devel] [PATCH proxmox-backup v2] etc: raise nofile soft limit for proxmox-backup-proxy

[Fabian Grünbichler]

On November 20, 2025 10:27 am, Christian Ebner wrote:
> Since commit 86d5d073 ("GC: fix race with chunk upload/insert on s3
> backends"), per-chunk file locks are acquired during phase 2 of
> garbage collection for datastores backed by s3 object stores. This
> however means that up to 1000 file locks might be held at once, which
> can result in the limit of open file handles to be reached.
> 
> Therefore, bump the nolimit for the proxmox-backup-proxy in the systemd
> service unit, while keeping the hard limit as defined in
> /etc/systemd/system.conf.
> 
> Signed-off-by: Christian Ebner <c.ebner at proxmox.com>
> ---
> Changes since version 1:
> - Bumpt the limit directly in the systemd service unit
> 
>  etc/proxmox-backup-proxy.service.in | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/etc/proxmox-backup-proxy.service.in b/etc/proxmox-backup-proxy.service.in
> index 7ca806aa4..52ad99356 100644
> --- a/etc/proxmox-backup-proxy.service.in
> +++ b/etc/proxmox-backup-proxy.service.in
> @@ -10,6 +10,7 @@ Type=notify
>  ExecStart=%LIBEXECDIR%/proxmox-backup/proxmox-backup-proxy
>  ExecReload=/bin/kill -HUP $MAINPID
>  PIDFile=/run/proxmox-backup/proxy.pid
> +LimitNOFILE=4096:524288

this change is not quite identical to the previous one though ;)
is 4k really enough? that means (not even) 4 GC runs, which could easily
happen with many datastores?

some sort of rationale why it's okay to do this would also be great,
after all there is a reason the default limit is so low..

>  Restart=on-failure
>  User=%PROXY_USER%
>  Group=%PROXY_USER%
> -- 
> 2.47.3
> 
> 
> 
> _______________________________________________
> pbs-devel mailing list
> pbs-devel at lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
> 
> 
>