3 Sep
2019
3 Sep
'19
11:53 a.m.
On 03.09.19 09:54, Juha Heinanen wrote:
Juha Heinanen writes:
Yesterday I started a trip and didn't get the time to look over the last
two traces, I should be able to do it later today.
In the first trace, the ping-pong looked like internally in the
libmysqlclient library before doing the acc insert, probably to wake
up/be sure connection is alive, and that seemed to be stuck.
Cheers,
Daniel
--
Daniel-Constantin Mierla -- www.asipto.com
www.twitter.com/miconda -- www.linkedin.com/in/miconda
Here is two traps taken about 8 minutes apart
during the same freeze:
https://box.tutpro.com/tmp/gdb_kamailio_20190902_120249
https://box.tutpro.com/tmp/gdb_kamailio_20190902_121032
During the freeze, it was possible to make MySQL queries to accounting
table using mysql cli command.
Looks like 8 UDP receivers are waiting for
futexlock and 8 are in
udp_rcv_loop. Those that are waiting for the lock are processing the
same reply. Same in both traps.
There is also mysql ping going on by timer process.