[RHEL6.2] system hangs when using mdraid
Issue
- 'tar' command involving an md raid device causes the machine to hang.
- Over the a recent two day period, there were 5 LoP LPAR hangs that required forced reboots to bring the LPARs back.
- No applications are IO intensive.
- The backup that hung the box was a single threaded tar operation.
- The same backup ran again successfully last night on another system.
- The systems that don't fail are making connections across the CAN to output the data to that system.
Environment
- Red Hat Enterprise Linux 6
- 2.6.32-220.el6.ppc64
- Hardware: IBM p780 (ppc64, 3864 Mhz, POWER7)
- Memory: 64 GB
- CPUs: 20
- Storage configuration: ext4 --> linear or snapshot LVM LV --> md0 (RAID1) --> sda|sdb --> ipr driver
- 'tar' command takes as input from an ext4 filesystem which is on an LVM snapshot of the root linear logical volume on top of the MD0 RAID1 device, and pipes the output to ssh. An log file is written to a unique directory on '/tmp/', which contains another ext4 filesystem which is a linear LVM logical volume coming from the same MD0 mirror.
- 'tar' command source filesystem
rootvg-root_snap dm-3 (253:3) (snapshot)
`- rootvg-root-real dm-0 (253:0) (linear)
`- md0 (9:0) (RAID1)
+---- sda ---- ipr (IBM Linux Power RAID) driver
|
+---- sdb ---- ipr driver
`- rootvg-root_snap-cow dm-2 (253:2) (linear)
`- md0 (9:0) (RAID1)
+---- sda ---- ipr (IBM Linux Power RAID) driver
|
+---- sdb ---- ipr driver
- 'tar' command output is piped to 'ssh'
- 'tar' command output log filesystem
rootvg-tmp dm-9 (253:9) (linear) [/tmp ext4 rw,relatime,barrier=1,data=ordered]
`- md0 (9:0)
`- md0 (9:0) (RAID1)
+---- sda ---- ipr (IBM Linux Power RAID) driver
|
+---- sdb ---- ipr driver
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase of over 48,000 articles and solutions.
Close
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.