CentOS Bug Tracker
Mantis Bug Tracker
Main | My View | View Issues | Roadmap
  

View Issue Details [ Jump to Notes ] [ Issue History ] [ Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0002379CentOS-5kernelpublic2007-10-07 18:022007-12-02 19:25
Reportermc 
PrioritynormalSeveritymajorReproducibilityalways
StatusresolvedResolutionfixed 
PlatformOSOS Version
Product Version5.0 - x86_64 
Target VersionFixed in Version5.1 
Summary0002379: problems mixing seagate 15k.4 and 15k.5 SAS drives under md with aic94xx
DescriptionOn an updated system (currently at 2.6.18-8.1.14.el5.centos.plus) i have problems during resync of an md raid5 array made up from 2 15k.4 and 1 15k.5 seagate drives.

Basically i created the raid and the initial resync went fine. After that, if i fail 1 disk, remove and readd (all through mdadm) the 15k.5, i get in my syslog:

Oct 7 15:20:59 cerberus kernel: md: using 128k window, over a total of 71143744
 blocks.
Oct 7 15:22:15 cerberus kernel: sd 0:0:0:0: SCSI error: return code = 0x0005000
0
Oct 7 15:22:15 cerberus kernel: end_request: I/O error, dev sda, sector 3734659
Oct 7 15:22:15 cerberus kernel: raid5: Disk failure on sda2, disabling device.
Operation continuing on 2 devices
Oct 7 15:22:15 cerberus kernel: sd 0:0:0:0: SCSI error: return code = 0x0005000
0
Oct 7 15:22:15 cerberus kernel: end_request: I/O error, dev sda, sector 3734923
Oct 7 15:22:15 cerberus kernel: sd 0:0:0:0: SCSI error: return code = 0x0005000
0
Oct 7 15:22:15 cerberus kernel: end_request: I/O error, dev sda, sector 3735171
Oct 7 15:22:15 cerberus kernel: sd 0:0:0:0: SCSI error: return code = 0x0005000
0
Oct 7 15:22:15 cerberus kernel: end_request: I/O error, dev sda, sector 3735427
Oct 7 15:22:15 cerberus kernel: sd 0:0:0:0: SCSI error: return code = 0x0005000
0

The disc otherwise is perfectly fine (passes vendor test suite) and has been already returned once. It also shows no problems at all, if hammered in a non raid config.

Disc firmware has been updated.

Sometimes, under linux rescue, during mdadm operations the system panics with the trace in attachment.
Additional InformationThis system ran centos 4.4, with md raid1 made up of the two 15k.4 just fine, without showing any problems during resync.
TagsNo tags attached.
Attached Files? file icon panic [^] (1,183 bytes) 2007-10-07 18:02 [Show Content]

- Relationships

-  Notes
(0006105)
mc (reporter)
2007-10-07 18:08

 Sorry. Forgot to say that sas controller is an onboard aic 9410w, from supermicro x7db3 motherboard.
(0006158)
mc (reporter)
2007-10-19 19:43
edited on: 2007-10-19 19:44

 This problem seems solved with 2.6.18-51 from http://people.redhat.com/linville/kernels/rhel5/ [^]
that carries some aic94xx fixes.

The same should apply to rhel/centos 5.1
(0006438)
mc (reporter)
2007-12-02 18:52

 Fixed with centos 5.1. Please close.

Thanks,

-m

- Issue History
Date Modified Username Field Change
2007-10-07 18:02 mc New Issue
2007-10-07 18:02 mc Status new => assigned
2007-10-07 18:02 mc File Added: panic
2007-10-07 18:08 mc Note Added: 0006105
2007-10-19 19:43 mc Note Added: 0006158
2007-10-19 19:44 mc Note Edited: 0006158
2007-12-02 18:52 mc Note Added: 0006438
2007-12-02 19:25 kbsingh@karan.org Status assigned => resolved
2007-12-02 19:25 kbsingh@karan.org Fixed in Version => 5.1
2007-12-02 19:25 kbsingh@karan.org Resolution open => fixed

Copyright © 2000 - 2011 MantisBT Group
Powered by Mantis Bugtracker