You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I get out of open file descriptors in case the icecast server is not available .
I see an additional used file descriptor on every reconnect to icecast, by default every 3 seconds.
If additionally a telnet socket is used by a client, the logs are flooded many times a second by 2018/04/28 13:19:57 [server:2] Failed to accept new client: "Too many open files in accept()"
until disk is full. This could be a separate issue. Maybe it could help to limit the output of this log line, for example print it out only once a second.
This has been tested with fresh installation using opam at ubuntu 16.04 LTS.
Thanks for this report, this looks pretty interesting. Any chance you could share more logs and/or more details on how to reproduce the issue. In particular, does the output.icecast reconnect a lot?
Hi,
I assigned a separate clock to output.icecast as described here:
https://github.com/savonet/liquidsoap/blob/master/doc/content/clocks.txt
I get out of open file descriptors in case the icecast server is not available .
I see an additional used file descriptor on every reconnect to icecast, by default every 3 seconds.
some time later...
some time later...
At some point you can find in log
2018/04/28 13:19:54 [/radio:2] Connection failed: could not initiate a new handler: Too many open files in socket()
the test script:
If additionally a telnet socket is used by a client, the logs are flooded many times a second by
2018/04/28 13:19:57 [server:2] Failed to accept new client: "Too many open files in accept()"
until disk is full. This could be a separate issue. Maybe it could help to limit the output of this log line, for example print it out only once a second.
This has been tested with fresh installation using opam at ubuntu 16.04 LTS.
Thanks & BR, Peter
The text was updated successfully, but these errors were encountered: