Date: Sat, 10 Nov 2012 19:29:19 -0500 (EST) From: Rick Macklem <rmacklem@uoguelph.ca> To: Xavier Beaudouin <kiwi@oav.net> Cc: freebsd-fs@freebsd.org Subject: Re: 9.0-RELEASE-p4 + NFS + ZFS = issues... :/ (probably a memory leak) Message-ID: <985721960.220092.1352593759104.JavaMail.root@erie.cs.uoguelph.ca> In-Reply-To: <B35D7BAE-303A-4765-9E62-E91A6BFCA65D@oav.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Xavier Beaudouin wrote:
> Hi Rick,
>
> I did several nfs mount using the udp patch you gave me and also TCP
> mount instead.
>
> Seem with vmstat -m that :
>
> NFS fh 1 1K - 342087 64
>
> Is going to growing and growing.
>
It is the first numeric field called "InUse" that indicates how many
are currently allocated. The above line indicates "1". Are you saying
that field keeps ratcheting up?
(342087 is just how many times a malloc of this type has happened. It
will normally keep increasing.)
For "vmstat -z", it is the 3rd numeric field called "INUSE" that would
keep increasing, if there was a leak.
If no NFS related items are ratcheting up for the fields mentioned above,
then there isn't a malloc() or uma_zalloc() leak.
--> Probably a ZFS issue and know nothing about ZFS. (I wouldn't know a
zpool, if it jumped up and bit me on the nose;-)
Hopefully some ZFS guys may jump in with suggestions on what to check, rick
> Umounting the nfs mount free some memory and I getting free memory
> again.
>
> Seems that :
>
> rsync -r -v -p -o -g --times --specials --delete --exclude=.snapshot
> /mnt/ /vol/hosting/
>
> is raising this issue on my own (on case B, see previous mails).
>
> Any another hints?
>
> /Xavier
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?985721960.220092.1352593759104.JavaMail.root>