[2656] in SIPB_Linux_Development
Re: bloom-picayune outages today
daemon@ATHENA.MIT.EDU (mhpower@MIT.EDU)
Tue Feb 16 16:26:26 1999
From: mhpower@MIT.EDU
Date: Tue, 16 Feb 1999 16:26:16 -0500
To: rtfm-maintainers@MIT.EDU
Cc: linux-dev@MIT.EDU, hartmans@MIT.EDU, sipb-afsreq@MIT.EDU
In-Reply-To: "[7403] in RTFM_Maintainers_Archive",
"[2614] in SIPB_Linux_Development",
"[3323] in SIPB-AFS-requests"
>Date: Sun, 14 Feb 1999 05:22:02 -0500
...
>The problem occurs when there's a lot of access to bloom-picayune via
>NFS from Linux machines running the 2.2 kernel.
As some of you noticed, there was one additional bloom-picayune outage
since this behavior was identified. The /redhat/5.2 directory on
bloom-picayune was NFS mounted from equal-rites on Monday at 18:55 --
bloom-picayune went down a few minutes after that (uname on
equal-rites lists a 2.2.0 kernel). So, unless someone has further
information about equal-rites at that time (I suppose it's possible,
although from what I can tell really unlikely, that equal-rites was
booted with a Linux 2.0.x kernel at that time on Monday), I still
believe the sole cause of all of the outages since February 11 has
been the inability of bloom-picayune's kernel to deal with NFS traffic
from machines running Linux 2.2 or similar kernels.
This also means I don't think there's any hardware problem with
bloom-picayune that's a direct cause of the outages.
Matt
