Help diagnosing slient PHP error

Discussion in 'General' started by stormy, Jan 4, 2014.

  1. stormy

    stormy Member

    I'm trying to debug a problem with WHMCS (domains/hosting billing script, now partially owned by cPanel).

    Some multiple domain registrations are failing silently, without anything being logged anywhere. It looks like the process simply dies or times out and leaves no trace. There's nothing in the PHP error_log, nor on the application own error logs.

    What can I enable to troubleshoot this? I would like to have as much error logging as possible.

    I would also like to increase any parameters that I can increase that could help with performance/avoiding timeouts. Any tips?
  2. NiteWave

    NiteWave Administrator

    this may help:

    lsws web admin -> Actions -> Toggle Debug Logging
  3. stormy

    stormy Member

    Thanks! Wow, this one really generates too much data! Unfortunately, I don't think I can leave it enabled for hours/days until the problem happens again.

    Here's the documentation for the Debug Logging in case someone stumbles into this thread:
    http://www.litespeedtech.com/support/wiki/doku.php?id=litespeed_wiki:toggle_debug_logging

    Are there any particular PHP parameters that would be different from a stock Apache cPanel installation, and that I could tweak? The server has plenty of CPU and RAM.
  4. NiteWave

    NiteWave Administrator

    >Some multiple domain registrations are failing silently, without anything being logged anywhere.

    then how do you know the failures, are there any entries in access log ?
  5. stormy

    stormy Member

    I can see the failures in the app itself. Some actions that should be carried out automatically are not done.

    This is all done via PHP, using cURL and APIs.
  6. stormy

    stormy Member

    Unfortunately I've had to pass on Litespeed for this server and go back to Apache.

    For whatever reason, Litespeed on this (high powered) VPS was causing issues with WHMCS.

    Besides the error I mentioned, I had another issue that vanished after the Apache switch. Sometimes pages would simply stop loading until I restarted the browser. I blamed it on Safari, although it was kind of strange.

    The problem is remarkably similar to the one we have with the API, it seems that Litespeed simply breaks the connections or holds them with no reply, or something.

    I wish I could help troubleshooting this but I can't. We have Litespeed on our dedicated servers and have found no issues at all, so I believe it's got something to do with this being a VPS.

    If for some reason it turns out not to be a Litespeed issue, I'll be sure to report back.

Share This Page