I have some "permanent" CIFS mounts, some from a Windows server and some from a Buffalo Terastation NAS.
If changes are made to either server, even just access rights, without umounting and re-mounting, both my Fedora servers "kernel panic" and freeze - most awkward if I'm off-site. Usually there's been no trace in the logs but today I have "BUG" mentioned by the Fedora-8 box:-
Sep 6 22:44:59 hawksvr5 smbd[12122]: Error writing 4 bytes to client. -1. (Co
nnection reset by peer)
Sep 6 22:47:02 hawksvr5 kernel: BUG: scheduling while atomic: lsof/12885/0x0003
0000
Sep 6 22:47:02 hawksvr5 kernel: Pid: 12885, comm: lsof Not tainted 2.6.25.14-69
.fc8 #1
Sep 6 22:47:02 hawksvr5 kernel: [<c04237ec>] __schedule_bug+0x4b/0x50
Sep 7 12:18:39 hawksvr5 rsyslogd: [origin software="rsyslogd" swVersion="2.0.2" x-pid="1971" x-info="http://www.rsyslog.com"][x-configInfo udpReception="No" udpPort="514" tcpReception="No" tcpPort="0"] restart
Sep 7 12:18:39 hawksvr5 kernel: rklogd 2.0.2, log source = /proc/kmsg started.
Sep 7 12:18:39 hawksvr5 kernel: Initializing cgroup subsys cpuset
Sep 7 12:18:39 hawksvr5 kernel: Linux version 2.6.25.14-69.fc8 (mockbuild@x86-6) (gcc version 4.1.2 20070925 (Red Hat 4.1.2-33)) #1 SMP Mon Aug 4 14:20:24 EDT 2008
Sep 7 12:18:39 hawksvr5 kernel: BIOS-provided physical RAM map:
While the Fedora-9 box, which also froze at 22:47 when I had changed access rights on the TeraStation, shows absolutely nothing (see below). At 22:47 I did "lsof | grep <sharename>" to see why I couldn't umount the CIFS shares on the FC8 box, having remembered to do so, while on the FC9 box, I only got as far as "df" to remind myself of the mounts. The latter had mention of CIFS on the complex "kernel panic" screen (no camera, I'm afraid) this afternoon, but a complete gap in the log:-
Sep 3 16:48:01 hawksvr6 mountd[15219]: authenticated mount request from 10.0.0.
178:835 for /backup (/backup)
Sep 5 09:37:56 hawksvr6 mountd[15219]: authenticated mount request from 10.0.0.
178:1012 for /backup (/backup)
Sep 5 12:40:12 hawksvr6 mountd[15219]: authenticated unmount request from 10.0.
0.178:780 for /backup (/backup)
Sep 7 12:19:19 hawksvr6 kernel: imklog 3.18.1, log source = /proc/kmsg started.
Sep 7 12:19:19 hawksvr6 rsyslogd: [origin software="rsyslogd" swVersion="3.18.1" x-pid="1723" x-info="http://www.rsyslog.com"] restart
Sep 7 12:19:19 hawksvr6 kernel: Initializing cgroup subsys cpuset
Sep 7 12:19:19 hawksvr6 kernel: Linux version 2.6.25.11-97.fc9.i686 (mockbuild@) (gcc version 4.3.0 20080428 (Red Hat 4.3.0-8) (GCC) ) #1 SMP Mon Jul 21 01:31:09 EDT 2008
Sep 7 12:19:19 hawksvr6 kernel: BIOS-provided physical RAM map:
Should I report this to the Fedora bugzilla?
I've searched it and can't see anything very relevant so far.
Anyone else experienced this and got workarounds?
Mike Yates CMBCS (ISSG)
IT Support Engineer
Hawkgrove Ltd - Software Systems Design
2, The Business Courtyard, Marl Pits Lane, Trudoxhill, Frome, Somerset, BA11 5DL, UK
+44 (0)1373 837900 fax: +44 (0)8700 518155
Registered in England: 2756481 VAT Reg: UK 601 1137 11
Registered Office: NSO Associates LLP, 75 Springfield Road, Chelmsford, Essex CM2 6JB
All e-mail is subject to contract and is not intended to create a legally binding agreement.
Hawkgrove Ltd will only be bound by an agreement in writing signed by an authorized signatory.
All outgoing email is scanned by Kerio, using McAffee Antivirus.
-- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
