Copy link to clipboard
Copied
Every once in a while we get the following errors in FIMS 4.5 edge log file and stream stop to work for the affected vhost until we restart FMS:
2014-11-07 19:40:59 28122 (i)2581537 Resending StartCore message to master (vhost: vhostnamehere, app: , inst: , tag: _2).
Server is an Intel Xeon E5405 running CentOS 5.1 with 16 GB RAM.
The error does not cause other vhosts to work just the affected one. Although the error seems to occur ar random times and not when server is under heavy load, here is a snapshot of the memory usage while the error is being reported and after FMS is restarted fixing the problem:
total used free shared buffers cached
Mem: 16434952 14240400 2194552 0 135128 7211612
-/+ buffers/cache: 6893660 9541292
Swap: 6088624 180 6088444
AFTER FMS restarted fixing issue:
total used free shared buffers cached
Mem: 16434952 10836248 5598704 0 182740 7716368
-/+ buffers/cache: 2937140 13497812
Swap: 6088624 180 6088444
Could it be an insufficient memory issue? I have noticed this error at least on one occasion when attempts were made to play a high bitrate (3000 kbps) VOD.
What could be causing this seemingly random issue and are there any tweaks that can be made to FMS to help prevent it? Or do we need to throw in more RAM? 16 GB seems like plenty enough.
Copy link to clipboard
Copied
Firstly 2014-11-07 19:40:59 28122 (i)2581537 Resending StartCore message to master (vhost: vhostnamehere, app: , inst: , tag: _2).
is an information and not an error
This message is emitted if a core process was requested to start but has not started for last over 30 seconds(determined by CoreTimeout in Server.xml).You may try to increase that time limit.
When a request is received by server and no core is available for that, the AMS would start a new core, and that request will be blocked till the time core is started(up and running).
That said, i admit i do not know why your core is getting stuck.