Intermittent Error 500

[JohnDar]

I have an ongoing intermittent issue which seems to happen at around the same time each week. My community fails to load, producing a generic 'Error 500' page. When this happens, checking with the ips4.php update checker works and returns a clean bill of health. The error happens on the standard theme and when logging into the ACP, as well as my modified theme.

The hosts have managed to fix the issue and have told me they have cleared the server cache and (once) told me it was a permissions issue. They can find nothing further amiss on their side, despite repeated support calls, each time it happens.

The system log produces the following entry when the issue occurs:

Quote

Error: Class '\IPS\Theme\class_forums_front_index' not found (0)
#0 /home2/projectd/public_html/applications/forums/modules/front/forums/index.php(216): IPS\_Theme->getTemplate('index')
#1 /home2/projectd/public_html/system/Dispatcher/Controller.php(101): IPS\forums\modules\front\forums\_index->manage()
#2 /home2/projectd/public_html/applications/forums/modules/front/forums/index.php(52): IPS\Dispatcher\_Controller->execute()
#3 /home2/projectd/public_html/system/Dispatcher/Dispatcher.php(153): IPS\forums\modules\front\forums\_index->execute()
#4 /home2/projectd/public_html/index.php(13): IPS\_Dispatcher->run()
#5 {main}

Are you able to shed any light on what might be causing this?

Thank you 🙂

[Marc Stridgen]

My suspicion here is that template files are occasionally not writing to disk, and you have disk template caching enabled. I would advise on switching this off if it is causing you issues. You can do this from System->Settiings->.Advanced Configuration, within your Admin CP 

[JohnDar]

Thank you for the suggestion, however, template caching is already disabled.

[Marc Stridgen]

What caches are they clearing to correct this? Is it server side caching you have running?

[kotaco]

I faced a similar issue with occasional errors, the cause of mine was some PHP resource settings. In particular "max_input_time" was set to the default of 30 seconds, it needed to be increased to avoid timeouts when writing files

Edited May 23, 2022 by kotaco

[Marc Stridgen]

Given this seems to be hosting related here, it may be worth giving the above a shot

[JohnDar]

Thank you, I will take a look at that, though the value is currently set to 60. May I ask what value fixed it for you?

3 hours ago, Marc Stridgen said:

What caches are they clearing to correct this? Is it server side caching you have running?

Yes, the hosts just told me they had cleared the server side cache. 

Edited May 23, 2022 by JohnDar

[Marc Stridgen]

I would advise on disabling the caching method being used, as it seems its clearly causing you an issue there. 

[JohnDar]

The hosts tell me they are not caching after all. Getting a straight answer about what they do when they investigate this is a challenge.

Having changed the "max_input_time" to 60, as suggested above, the errors have become less frequent and seem to last around three minutes.

On the most recent occasion, the usual 500 Error was intermittently replaced with this:

Note that this was using the default theme and also logging into the ACP. I don't know if this is at all significant, but thought I'd post it, just in case.

[Marc Stridgen]

Please persist with your hosting company to find out what they actually do when they resolve the issue, as this may be the key to understanding what is going wrong

[JohnDar]

It would appear that the hosts weren't  actually affecting the issue. The Forum was coming back after a few minutes on its own. I suspect they thought they had fixed it, each time they started investigating.

Following the advice above from @kotaco, to increase the max_input_time, I found that increasing from 30 seconds to 60 seconds cut the errors by about 70%. I then increased this to 90 seconds and a week in, the errors have stopped.

Result.

I don't know what changed that required this modification. I have always run on the default 30s limit, previously. It's difficult to determine whether this issue started following an IPS software update or whether it might be related to some change on the hosting server.

 

Edited June 14, 2022 by JohnDar

[Jim M]

10 minutes ago, JohnDar said:

I don't know what changed that required this modification. I have always run on the default 30s limit, previously. It's difficult to determine whether this issue started following an IPS software update or whether it might be related to some change on the hosting server.

If it is indeed a write issue, as kotaco suggested, this would be something to bring up to your hosting provider. This would be solely impacted by their server configuration and drive performance.

[Marc Stridgen]

In addition to this, check your server error logs at times just before this starts to happen. Its likely you will see what is the underlying cause of the problem there