I didn't get to the logs yet, but with this change added I see many do_install tasks failing with exit code '134'.

It might have different cause, but I wasn't seeing this after last oe-core upgrade before this last pseudo SRCREV bump.


There isn't temp/log.do_install* (for whatever reason) and in pseudo.log I see following 3 lines being repeated couple times:

debug_logfile: fd 2
pid 47520 [parent 47518], doing new pid setup and server start
Setup complete, sending SIGUSR1 to pid 47518.

250 times even for very simple recipe:
sysvinit-inittab/2.88dsf-r10$ grep -c "Setup complete, sending SIGUSR1" pseudo/pseudo.log
250


On Fri, Feb 16, 2018 at 6:51 PM, Seebs <seebs@seebs.net> wrote:
On Fri, 16 Feb 2018 17:55:54 +0100
Martin Jansa <martin.jansa@gmail.com> wrote:

> $ grep -c "tried to close client 0 (highest is 1)"
> update-rc.d/0.7-r5/pseudo/pseudo.log
> 579655922
>
> All pseudo processes I've seen on various servers which were building
> with this change included got stuck until disk space run out..

Oh-hoh! That could be the epoll thing that I wasn't able to reproduce.

> commit 26e30fa2e1a0fe4e885d7eea3f55d23cd2c3158f
> Author: Seebs <seebs@seebs.net>
> Date:   Fri Feb 16 11:49:49 2018 -0600
>
>     Allow closing client 0

Added a thing to master. It looks like the failure mode is that a
client ended up on fd 0, and for some reason, back in 2010, I thought
that would be invalid and did not allow it. So change is to replace "<=
0" with "< 0".

-s