[Bug 271945] panic: deadlres_td_sleep_q: possible deadlock detected for 0xfffffe0133324ac0 (stat), blocked for 1801328 ticks

From: <bugzilla-noreply_at_freebsd.org>
Date: Wed, 12 Jul 2023 03:05:06 UTC

--- Comment #2 from Cy Schubert <cy@FreeBSD.org> ---
I think the common thread here is poudriere-bulk is running. When running with
-j number of jobs the likelihood of this occurring is > 50%. Whereas with -j 1
it will never happen.

I have no zpools with errors.

I have another machine, not quite configured the same (fewer disks), also
running poudriere but in this case the poudriere jails are all i386 jails
whereas the machine that panics, the poudriere jails are all amd64. I don't
know if this has a bearing on this problem or not.

My panic string is usually like this, though at times I will see a c++ process
in a detected deadlock.

Unread portion of the kernel message buffer:
panic: deadlres_td_sleep_q: possible deadlock detected for 0xfffffe00c0d6d740
ostgres), blocked for 180239 ticks

cpuid = 0
time = 1689121195
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe00851b7e00
vpanic() at vpanic+0x149/frame 0xfffffe00851b7e50
panic() at panic+0x43/frame 0xfffffe00851b7eb0
deadlkres() at deadlkres+0x32a/frame 0xfffffe00851b7ef0
fork_exit() at fork_exit+0x82/frame 0xfffffe00851b7f30
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00851b7f30
--- trap 0, rip = 0, rsp = 0, rbp = 0 ---
Uptime: 8h51m9s

It's usually postgresql which is the DBMS for gitea. But, as intimated above,
sometimes it's a c++ that's listed as the "cause."

You are receiving this mail because:
You are the assignee for the bug.