Thread (7 messages) 7 messages, 3 authors, 2015-10-23

Re: Hung RAID5 array with discard

From: Neil Brown <hidden>
Date: 2015-10-23 05:51:59

Peter Kieser [off-list ref] writes:
On 2015-10-22 9:42 PM, Neil Brown wrote:
quoted
Peter Kieser [off-list ref] writes:
quoted
FYI I ran into this problem in 3.18.22 when forcing RAID5 TRIM support
w/ raid456.devices_handle_discard_safely on Intel DC S3500 SSDs.
Can you please be precise about the problem you experienced.
Complete kernel messages are a minimum.
Anything interesting that might have been happening at the time might
help.

Thanks,
NeilBrown
Same behaviour as the original poster. Enabled the module knob, then ran 
mkfs.ext4, which starts to do "Discarding device blocks" and then the 
machine hard locked. No kernel messages.
Thanks.
The original poster reported a hung-task dump...
When you say "hard locked" - can you still access from another
window/login, or is it totally forzen?
If former, are then any processes in D start?  maybe mdXX-raid5?
Can you get /proc/$PID/stack of that process?

If totally frozen - do you have a console?  Can you alt-sysrq-W ??

Or have you since rebooted and the problem isn't repeatable?

Thanks,
NeilBrown


Attachments

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help