Bug 167445 - multipathd sometimes dies when many disks being added by multipath
Summary: multipathd sometimes dies when many disks being added by multipath
Alias: None
Product: Red Hat Enterprise Linux 4
Classification: Red Hat
Component: device-mapper-multipath
Version: 4.0
Hardware: i386
OS: Linux
Target Milestone: ---
: ---
Assignee: Ben Marzinski
QA Contact:
Depends On:
TreeView+ depends on / blocked
Reported: 2005-09-02 17:53 UTC by Stoo Davies
Modified: 2010-01-12 02:21 UTC (History)
2 users (show)

Clone Of:
Last Closed: 2005-10-11 04:16:48 UTC

Attachments (Terms of Use)

Description Stoo Davies 2005-09-02 17:53:32 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.9) Gecko/20050715 Firefox/1.0.5

Description of problem:
I'm doing some initial testing on RHES4-U2 beta with 3PAR InServ storage arrays and dm-multipath/multipath-tools.
The devices on the storage array were already partitioned and had filesystems on them, they were discovered at the scsi layer by reloading the qla2300 module, and then multipath was run to create the /dev/mapper/xxxx entries.
At this point the CPU load hit 100% on the box, the process that was stuck was..
/sbin/mpath_ctl "add map dm-2"
after looking at it some more it seems that multipathd had died during the setup of these maps.   I assume the mpath_ctl process was stuck waiting for a response from multipathd.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
See desc

Additional info:

Comment 1 Alasdair Kergon 2005-10-03 22:06:46 UTC
Exactly which version of the multipath package was this with?
(Fixes were applied for something similar to this.)

Comment 2 Alasdair Kergon 2005-10-03 22:08:47 UTC
The daemon's read/write loops still need fixing to bring them into line with the

Comment 3 Stoo Davies 2005-10-03 23:04:47 UTC
It was device-mapper-multipath-0.4.5-5.2.RHEL4

Comment 4 Alasdair Kergon 2005-10-04 13:14:52 UTC
Please retest with -6.0.RHEL4 then: hopefully you'll find this particular
problem is fixed.

Comment 5 Stoo Davies 2005-10-10 17:08:14 UTC
Looks good, didn't see anything hang over the weekends test runs.  Thanks

Comment 6 Ben Marzinski 2005-10-11 04:16:48 UTC
I'm closing the bug. Let me know if you see this again.

Note You need to log in before you can comment on or make changes to this bug.