Port-xen archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Disk-related hangs since 11.0 upgrade



On 11.10 11:15, Manuel Bouyer wrote:
On Sun, Oct 11, 2026 at 11:08:52AM +0300, Staffan Thomen wrote:
On 11.10 11:06, Manuel Bouyer wrote:
On Sun, Oct 11, 2026 at 10:55:04AM +0300, Staffan Thomen wrote:
As I can access the SOL console and ddb, but don't know how to find
out what's hanging I would appreciate some suggestions for when this
happens next.

If you can enter ddb you could use 'ps' to list processes and their
wait channels.


Yep, got as far as running ps. There was a lot there; what should I be
looking for?

Hard to tell ... one starting point could be looking at a process
which you known is hung with ps/a. This will give you its proc structure
address. Then you can get a stack trace with tr/a

Another way, if you see lots of processes waiting on tstile,
would be to start with show all tstiles. Once you found the lwp which
holds the lock causing the hang, get a stack trace of it.


Following up on this, I noticed the system being hung while I was away today but fumbled around a bit in ddb remotely via the SOL console before rebooting.

My one known hung process was getty (I tried to log in on the console) and it was stuck in biowait. I have that backtrace, but it seems fairly unintresting; ufs_blkatoff -> breadn -> biowait -> cv_wait -> sleepq_block

In hindsight I should have looked for more processes in biowait and gotten their traces... Maybe next time I will have more presence of mind.

It certainly still seems disk-related, and I'm wondering about the hardware again.

There were no processes in tstile that I could see.

Staffan

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature



Home | Main Index | Thread Index | Old Index