--- Begin Message ---
- To: Debian Bug Tracking System <submit@bugs.debian.org>
- Subject: open-iscsi: Logging in causes seg fault in udevd and server crash
- From: Andrew Moise <chops@demiurgestudios.com>
- Date: Thu, 21 Feb 2008 14:00:32 -0500
- Message-id: <20080221190032.3359.96124.reportbug@localhost.localdomain>
Package: open-iscsi
Version: 2.0.730-1etch1
Severity: important
I attempted to use open-iscsi to log in to an iSCSI target (a Dell
MD3000i) attached to my system, and when I logged in the server became
unresponsive from the network and console. When I rebooted and tried
again from the console, I was able to see the symptoms of the crash
before the system became unresponsive. This was the error (retyped from
a camera-phone snapshot):
moise@breakout:~$ sudo iscsiadm --mode node --targetname iqn.1984-05.com.dell:powervault.6001e4f0001997780000000047b6abe6 --portal 10.101.152.12:3260 --login
scsi1 : iSCSI Initiator over TCP/IP
Vendor: DELL Model: MD3000i Rev: 0670
Type: Direct-Access ANSI SCSI revision:
udevd[3477] general protection rip:2b3df16cadca rsp:7fffb9689d08 error:0
udevd[1426] general protection rip:4044ccf rsp:7fffb9693cb0 error:0
Here are all the relevant log entries I could find from the session
where this crash occurred. The interface with the iSCSI target
connected to it was _not_ up when the session started, so I upped the
interface and then attempted to login again.
Feb 21 13:08:16 localhost kernel: Loading iSCSI transport class v1.1-646.<5>iscsi: registered transport (tcp)
Feb 21 13:08:16 localhost kernel: iscsi: registered transport (iser)
Feb 21 13:08:16 localhost iscsid: iSCSI logger with pid=2966 started!
Feb 21 13:08:17 localhost iscsid: transport class version 1.1-646.iscsid version 2.0-730
Feb 21 13:08:17 localhost iscsid: an InitiatorAlias= is required, but was not found in /etc/initiatorname.iscsi
Feb 21 13:08:17 localhost iscsid: iSCSI daemon with pid=2967 started!
Feb 21 13:08:17 localhost iscsid: iSCSI sync pid=2968 started
...
Feb 21 13:13:23 localhost iscsid: connect failed (101)
... then up the interface and try again:
Feb 21 13:13:57 localhost kernel: bnx2: eth0: using MSI
Feb 21 13:13:57 localhost kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready
Feb 21 13:13:57 localhost ntpdate[3455]: the NTP socket is in use, exiting
Feb 21 13:14:00 localhost kernel: bnx2: eth0 NIC Link is Up, 1000 Mbps full duplex, receive & transmit flow control ON
Feb 21 13:14:00 localhost kernel: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Feb 21 13:14:10 localhost kernel: eth0: no IPv6 routers present
... and that's the last thing that makes it into the log on disk
before the crash.
I'm going to keep investigating (including trying to log in from a sid
system); if I learn anything more about specific elements of the
configuration that may have caused this error, I'll send them along.
This is a production server, though, so I'll have to tread a little
carefully, and I'll certainly have to wait until some downtime is
available before continuing to investigate on this server.
Cheers!
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: amd64 (x86_64)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-6-amd64
Locale: LANG=en_US.ISO-8859-1, LC_CTYPE=en_US.ISO-8859-1 (charmap=ISO-8859-1)
Versions of packages open-iscsi depends on:
ii libc6 2.3.6.ds1-13etch5 GNU C Library: Shared libraries
open-iscsi recommends no packages.
-- no debconf information
--- End Message ---