Okay, just got distracted by this happening again. The way I'm reproducing it
is by ^C'ing a Cassandane run (which cleanly shuts down the cyrus instances it
had started) during the JMAP tests, and I guess I'm sometimes getting lucky on
the timing and hitting this.
On Wed, Dec 4, 2019, at 2:22
So, using my strace output from the other week as an example:
> pselect6(13, [8 9 11 12], NULL, NULL, NULL, {[], 8}) = 1 (in [11])
The arguments here are:
* nfds: 13 (12+1)
* readfds: [8 9 11 12]
* writefds: NULL
* exceptfds: NULL
* timeout: NULL
* sigmask: {[], 8}
The interesting bits are:
Hello Ellie,
this is exactly what I see (countless pselect calls), but I have as second
parameter of pselect a much larger array. I
just observed that on killing master, it terminates all cyrus processes but two
(httpd and notifyd). Then I try to
connect to that processes (gdb). This does
Saw something similar just now when I killed a cassandane run off prematurely.
One cyrus master process wound up spinning like this:
pselect6(13, [8 9 11 12], NULL, NULL, NULL, {[], 8}) = 1 (in [11])
pselect6(13, [8 9 11 12], NULL, NULL, NULL, {[], 8}) = 1 (in [11])
pselect6(13, [8 9 11 12],
Can you strace the master process next time it's spinning at 100%? What is it
doing at that time?
On Tue, Nov 26, 2019, at 1:29 AM, Дилян Палаузов wrote:
> Hello,
>
> > I run cyrus imap 3.0.x with some private changes.
> >
> > Sometimes when stop the master process, the master process
Hello,
> I run cyrus imap 3.0.x with some private changes.
>
> Sometimes when stop the master process, the master process utilizes one CPU
> core to 100% for 5 minutes. After the fifth
> minute, systemd enforces kill -9. When I attach to the maste process, I see
> that it some janitor does
