ndemou Posted October 15, 2018 Report Share Posted October 15, 2018 We had problems with calls not connecting or getting no audio today and when I examined the logs I show many serious errors like these:  ... Could not open cdr....bin for writing  ... Port ...: Could not allocate sockets!  ... UDP(IPv4): Could not open socket (EMFILE) lsof | grep ^pbxctrl | wc -l showed more that 8160 lines. Looking at the output of lsof I noticed that 99% of the lines are like this: COMMAND PID TID USER FD TYPE DEVICE SIZE/OFF NODE NAME pbxctrl 1280 1307 root 1005u IPv6 10107481 0t0 UDP *:49689 I wonder if so many open files(UDP sockets) is normal. It doesn't seem to be so I restarted pbxctrl and the problems went away (after a painful 5 minutes). I'm monitoring the output of `lsof | grep ^pbxctrl | wc -l` and it shows me anywhere from 1200 to 1500 open files (with moderate traffic). I've raised my soft & hard limit of open files from 1024 & 4096 to 4000 & 8000 with prlimit -n -p`pidof pbxctrl` --nofile=4000:8000 Some basic facts about my system: Software-Version: 60.0.3 (CentOS32) Build Date: May 1 2018 04:42:47 uname -a Linux ... x86_64 x86_64 x86_64 GNU/Linux Questions: Have we hit a bug or is such an amount of open sockets normal? If it's normal what's the recommended configuration for CentOS? Is the 5 minutes to restart normal? How could we cut it down? Is it OK that I'm running 64bit CentOS but pbxctrl reports CentOS32? Quote Link to comment Share on other sites More sharing options...
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.