Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 30 Apr 2003 23:38:52 -0500
From:      Dan Nelson <dnelson@allantgroup.com>
To:        "Dave [Hawk-Systems]" <dave@hawk-systems.com>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: kernel: looutput: mbuf allocation failed
Message-ID:  <20030501043851.GE90046@dan.emsphone.com>
In-Reply-To: <DBEIKNMKGOBGNDHAAKGNCEPANGAB.dave@hawk-systems.com>
References:  <3EB0321A.5020106@potentialtech.com> <DBEIKNMKGOBGNDHAAKGNCEPANGAB.dave@hawk-systems.com>

next in thread | previous in thread | raw e-mail | index | archive | help
In the last episode (Apr 30), Dave [Hawk-Systems] said:
> >> Recently had a box go dark for a few minutes...  pulled the
> >following from the
> >> logs shortly after the box revived itself.
> >>
> >> 	/kernel: looutput: mbuf allocation failed
> >>
> >> been a while since we have touched the kernel config for this
> >> server...  any pointers as to what settings we should lookat or
> >> touch to prevent this nastiness from occuring again?
> >
> >What does 'netstat -m' have to say?
> 
> nx1# netstat -m
> 1045/1856/6144 mbufs in use (current/peak/max):
>         977 mbufs allocated to data
>         68 mbufs allocated to packet headers
> 970/1536/1536 mbuf clusters in use (current/peak/max)
> 3536 Kbytes allocated to network (76% of mb_map in use)
> 27349 requests for memory denied
> 14 requests for memory delayed
> 0 calls to protocol drain routines

You've already maxed out your mbuf clusters, and possibly dropped 27349
packets.  Try putting "kern.ipc.nmbclusters=4096" in /boot/loader.conf
and rebooting.  1536 seems really low for a web/DNS server.
 
-- 
	Dan Nelson
	dnelson@allantgroup.com



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030501043851.GE90046>