2.2.2 Bug? RPC: sendmsg returned error 101

Jan Sevelin (fjs@mc.hik.se)
Mon, 8 Mar 1999 11:05:11 +0100 (MET)


Symptom: Under kernel 2.2.2 when shutting down networking in runlevel 6
the shutdown process gets stuck repeating "RPC: sendmsg returned error 101"

Umounting or not umounting nfs filesystems before this point seem to make no
difference, nor does it seem to matter if these filesystems have been
mounted with nolock or not.

If the mounts are hard the shutdown doesn't get further without some
slight violence. If the mounts are soft the servers will reboot, but
will not for some reason umount local filesystems nor stop the
RAID devices. Not to graceful.

We have two pairs of boxes running 2.2.2. The pair that is affected differs
from the unaffected pair in that they have SMP and MD enabled (with
define MAX_MD_DEV 6). The problem is invariably reproduced on both the
affected boxes.

The systems are RedHat 5.2, with apropriate upgrades. The nfs stuff is
2.2beta37.

I think the bug is real. I'm not certain where it lives. I have temporarily
solved the problem for us by patching net/sunrpc/xprt.c but this is deep
water for me. I hack perl, but as long as the kernel is not rewritten in that
eminent language I rather keep my hands away... :-) I would appriciate a
more trustworthy fix, or a bless of my fix by someone more trustworty in
these matters than I... :-)

*** xprt.c Wed Jan 20 22:44:53 1999
--- ../../../linux-2.2.2-jan/net/sunrpc/xprt.c Mon Mar 8 10:58:36 1999
***************
*** 222,223 ****
--- 222,225 ----
switch (result) {
+ case -ENETUNREACH:
+ break;
case -ECONNREFUSED:

Regards

/Jan Sevelin
Dpt of Media, Communication & Journalism, University of Kalmar

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/