No resolution to this point.
Original question:
I have had a nagging problem for the last few months. I have looked at SunSolve
and applied the recommended patches. I am still having the problem on one
particular system.
About every week or two, one of my SPARCstation 10's running 4.1.3 will cease to function
(i.e. RIP in perfmeter). I look at the console and one of the last messages is "mbuf map full".
I have a login prompt at the console, but once I type in "root" and hit <Return>, it hangs.
All of my other servers complain that they can't speak NFS with the
hung server. My only recourse so far has been to <L1><A> the system, "sync" and then
"reset". It usually reboots fine and plays nice for another week or two.
Has anyone else solved this problem? Is there some device on my network which may be
causing the server to run out of mbufs?
Note: The problem seemed to start after I added FDDI interfaces to all of my servers.
Thanks to:
White Gary SrA USAFE CSS/SCOE
Juan Iturbe
Caleb Warner
Randi Simons
Art Schoenstadt
-------------------
I thing Gary's response may be part of the answer I am looking for:
I seem to remember having this problem some time ago on several machines
which were running FDDI. I think the solution was to change the rsize
and wsize in the vfstab or automount maps. We also were getting NFS
time-outs at the same time. I can't recall what we changed the rsize
and wsize to, it's been several years and I am not in a new job and
don't have access to those systems to look. I hope this helps.
-----------------------
Does anyone have anything to add to Gary's answer?
Thanks.
Steve.Kirkpatrick@andataco.com
Network Systems Manager
ANDATACO, San Diego, CA USA
http://www.andataco.com
(619) 646-1550 (Voice)
(619) 453-5793 (FAX)
1-800-SKY-8888, PIN: 1288059 (Pager)
p_stepkirk@andataco.com (Pagemail)
"The great thing about owning a basset hound is that no matter how bad you
feel, they always look worse."
--- Lewis Grizzard
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:56 CDT