Bug 721345

Summary: Boot from multipath device on IBM SVC fails for kernel 2.6.18-238.12.1.el5.x86_64
Product: Red Hat Enterprise Linux 5 Reporter: Paweł Brodacki <ofbugsandmen>
Component: kernelAssignee: Red Hat Kernel Manager <kernel-mgr>
Status: CLOSED NOTABUG QA Contact: Red Hat Kernel QE team <kernel-qe>
Severity: low Docs Contact:
Priority: unspecified    
Version: 5.6CC: jfeeney
Target Milestone: rcKeywords: Regression
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-08-18 13:54:08 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
multipath.conf from an affected server
none
Working multipath.conf none

Description Paweł Brodacki 2011-07-14 10:52:58 UTC 
Created attachment 512856 [details]
multipath.conf from an affected server

Description of problem:
RHEL 5 Update 6 x86_64 system does not boot with kernel 2.6.18-238.12.1.el5.x86_64


Version-Release number of selected component (if applicable):
kernel-2.6.18-238.12.1.el5.x86_64
device-mapper-multipath-0.4.9-41.el6.x86_64

How reproducible:
Every time.

Steps to Reproduce:
1. Install RHEL 5.6, root filesystem on LVM on a multipath device from IBM SVC, 
   boot partition on /dev/mapper/mpath0p1.
2. Upgrade device-mapper-multipath to 0.4.9-41.el6.x86_64.
2. Install kernel-2.6.18-238.12.1.el5.x86_64.rpm
3. Reboot and boot with the new kernel.
  
Actual results:
Multipathing does not work, LVM complains about duplicate PVID, and boot process stops and Checking filesystems, because fsck.ext3 cannot find boot partition, because /dev/mapper/mpath0p1 does not exist.

Expected results:
System boot normally.

Additional info:
Re-creation of initrd with

# mkinitrd /boot/initrd-2.6.18-238.12.1.el5.img 2.6.18-238.12.1.el5

does not solve the problem.

Installation of kernel-2.6.18-269.el5.x86_64.rpm from http://people.redhat.com/jwilson/el5/269.el5/x86_64/ solves the problem.
Comment 1 Paweł Brodacki 2011-07-15 09:51:14 UTC 
Created attachment 513349 [details]
Working multipath.conf

Correct multipath conf.
Comment 2 Paweł Brodacki 2011-07-15 09:55:59 UTC 
The problem was solved by modification of multipath.conf (attached as Working multipath.conf).

The original multipath.conf blacklisted all devices via devnode and wwid, then added exception via device section. According to documentation of DM Multipath (p.13):
"When specifying devices in the blacklist_exceptions section of the configuration file, you must specify the exceptions in the same way they were specified in the blacklist. For example, a WWID exception will not apply to devices specified by a devnode blacklist entry, even if the blacklisted device is associated with that WWID. Similarly, devnode exceptions apply only to devnode entries, and device exceptions apply only to device entries."

The original file yielded:
o) A bootable system with blacklisted local drives for kernels 2.6.18-238.el5 and 2.6.18-269.el5, which is what was intended, but is incorrect behaviour according to documentation.
o) A non-bootable system for kernel 2.6.18-238.12.1.el5, which is the proper behaviour according to the documentation.

PEBKAC, I'm afraid, sorry for the noise.
Comment 3 John Feeney 2011-08-18 13:54:08 UTC 
Closing per comment #2.