[pve-devel] memory leak in PVE/QMPClient.pm

[Dietmar Maurer]

OK - fixed. Please test again.

> -----Original Message-----
> From: Stefan Priebe - Profihost AG [mailto:s.priebe at profihost.ag]
> Sent: Montag, 29. Oktober 2012 13:05
> To: Dietmar Maurer
> Cc: Alexandre DERUMIER; pve-devel at pve.proxmox.com
> Subject: Re: [pve-devel] memory leak in PVE/QMPClient.pm
> 
> i would suggest the following:
> 
> diff --git a/bin/pvestatd b/bin/pvestatd index cb0d5cd..0345fe8 100755
> --- a/bin/pvestatd
> +++ b/bin/pvestatd
> @@ -384,7 +384,7 @@ for (;;) { # forever
> 
>          my $mem = PVE::ProcFSTools::read_memory_usage();
> 
> -       if (!defined($initial_memory_usage)) {
> +       if (!defined($initial_memory_usage) || $cycle < 10) {
>              $initial_memory_usage = $mem->{resident};
>          } else {
>              my $diff = $mem->{resident} - $initial_memory_usage;
> 
> Mit freundlichen Grüßen
>    Stefan Priebe
> Bachelor of Science in Computer Science (BSCS) Vorstand (CTO)
> 
> -------------------------------
> Profihost AG
> Am Mittelfelde 29
> 30519 Hannover
> Deutschland
> 
> Tel.: +49 (511) 5151 8181     | Fax.: +49 (511) 5151 8282
> URL: http://www.profihost.com | E-Mail: info at profihost.com
> 
> Sitz der Gesellschaft: Hannover, USt-IdNr. DE813460827
> Registergericht: Amtsgericht Hannover, Register-Nr.: HRB 202350
> Vorstand: Cristoph Bluhm, Sebastian Bluhm, Stefan Priebe
> Aufsichtsrat: Prof. Dr. iur. Winfried Huck (Vorsitzender)
> 
> Am 29.10.2012 13:01, schrieb Stefan Priebe - Profihost AG:
> > Am 29.10.2012 12:57, schrieb Dietmar Maurer:
> >> fixed - please test again (sorry)
> >
> > My test was too short. Does not work to me:
> >
> > Oct 29 12:59:09 cloud1-1202 pvestatd[5868]: starting server Oct 29
> > 12:59:10 cloud1-1202 pvestatd[5868]: restarting server after 2 cycles
> > to reduce memory usage (free 35692544 (7036928) bytes) Oct 29 12:59:10
> > cloud1-1202 pvestatd[5868]: server shutdown (restart) Oct 29 12:59:11
> > cloud1-1202 pvestatd[5868]: restarting server Oct 29 12:59:21
> > cloud1-1202 pvestatd[5868]: restarting server after 2 cycles to reduce
> > memory usage (free 37605376 (5599232) bytes) Oct 29 12:59:21
> > cloud1-1202 pvestatd[5868]: server shutdown (restart) Oct 29 12:59:21
> > cloud1-1202 pvestatd[5868]: restarting server Oct 29 12:59:31
> > cloud1-1202 pvestatd[5868]: restarting server after 2 cycles to reduce
> > memory usage (free 37605376 (5595136) bytes) Oct 29 12:59:31
> > cloud1-1202 pvestatd[5868]: server shutdown (restart) Oct 29 12:59:32
> > cloud1-1202 pvestatd[5868]: restarting server Oct 29 12:59:42
> > cloud1-1202 pvestatd[5868]: restarting server after 2 cycles to reduce
> > memory usage (free 37605376 (5599232) bytes) Oct 29 12:59:42
> > cloud1-1202 pvestatd[5868]: server shutdown (restart) Oct 29 12:59:42
> > cloud1-1202 pvestatd[5868]: restarting server
> >
> > I've seen that it takes up to 10 cycles to have a constant memory size.
> > I think this is due to the fact that perls garbage collection is not
> > running every after each cycle.
> >
> > Stefan