Failed to start one instance.

[sahith]

I have been using litespeed webserver from past 3 years, but never faced these errors. From the past few weeks, occasionally some websites are going down and showing 503 error pages. with errors in log showing as

2017-01-03 10:37:57.626 [ERROR] [APVH_user_Suphp:]: Failed to start one instance.

I just want to know why these errors are happening, and why litespeed not restarting itself in these circumstances, until i check websites manually, i have no idea that they are down.

 

[NiteWave]

is cloudlinux running ? maybe LVE limit too low when on busy time.
please check and adjust LVE:
Concurrent connections(EP) and/or
Number of processes(nPROC)

 

[sahith]

Not using Cloudlinux, just Centos and Cpanel. And using csf firewall, and it is blocking 99% attacks just fine, So there is no chance of attack too.

 

[NiteWave]

I think the error is related to "PHP suEXEC Max Conn" setting

1. increase it
2. watch real-time stats, when the error happen, see if WaitQ number is big (normally it's 0)
3. adjust lsws web admin->Server->Dynamic Requests/second = 1, may mitigate the error number

 

[sahith]

My setting of Auto Fix 503 Error was set to 'NO', now i changed to to "YES", i might have to wait to see how this behaves when sites throwing 503 errors. It says it only restarts the litespeed when there are more than 30 503 errors in the span of 30 seconds,

Is there a way to change that config to like restart litespeed when there are more than 30 503 errors in the span of 60 seconds?

 

[Pong]

I believe you should focus on fixing your 503 PHP error. "Auto Fix 503 Error" is only a tmp workround, but you have to fix the root cause of php crash.

check error log and stderr log. also system log to find out the root cause.

Please follow this guild:
https://www.litespeedtech.com/support/wiki/doku.php/litespeed_wiki:php:503-errors

expecially check if lsphp was killed by CFS.