My sites are all DOWN - 500 Internal Server Errors

[mschrier]

I submitted a ticket, but it's been well over 24 hours and no response from Support.  All of our hosted websites are reporting 500 Internal Server errors or similar.

The error logs are full of the following type of message:

[error] [client 34.229.252.252] (12)Cannot allocate memory: couldn't create child process: /opt/suphp/sbin/suphp for /home/XXX/public_html/XXX.XXX/index.php, referer: XXX

We did not make any changes - in fact I rarely visit these sites and made no edits since October or November.  I don't know how long this problem has been occurring.

Anything I can check on from my end??

[MrPhil]

That's an unusual 500 error, so it will probably take support to fix it. The only thing I can think of at the moment is that you check your .htaccess and php.ini files to see if they're still there, and readable, and they contain your settings (especially any for allowed memory size). Settings that haven't been touched in years and use obsolete formats (e.g., php_value and php_flag in .htaccess) can cause 500 errors, but it's usually because the command is no longer permitted in that file. Anyway, make sure such PHP settings are in your php.ini file and not .htaccess. If that doesn't fix it, support will have to step in.

You got an automated response to your ticket, which tells you they've received it? Also, a problem like this may be more complex than run of the mill problems, and could take more time to deal with.

[youfoundjake]

mschrier..
I had a smiliar issue, and per supports answer on my ticket...

"Since your server has been switched to a new hardware machine which is under the newer version of operating system (CentOS 7), dual version of PHP (5.6 & 7) and MySQL 5.6, some of parameter settings in .htaccess are no longer with the new environments such as 'SetEnv'. You just need to remove or comment the coding line which has 'SetEnv' so that your website should back to be working."

take a look at your php.ini file...

[josephbe]

My memory utilization has been very high on these new servers due to stuck processes.

A week ago I had a DB table lock up and become completely inaccessible.  6 PHP requests for the same page that query that table got stuck and each consumed approx 1/6th the memory available on my account, and stayed that way for days.  All pages/domains were giving 500s.  The error log was full of 'unable to allocate memory.'  I spoke with support 14 times over 3 days to kill processes and revive the server and track down what was going on.  I reloaded the bad table to another name to restore functionality.  Days later, completely independent of either existing ticket, someone there created a new ticket about a stuck MySQL process and then killed the initial process that was locking the DB table (thank you!).

Fast forward a week.  Again I'm getting 500s across my sites.  I've spoken to support 4 times in the past 24 hours.  I have a ticket that has been raised to the highest urgency, and still has not been commented on for 20 hours.  In the past week the phone support reps have lost the ability to view and/or kill processes on the server.  I don't see a way to even view what pages FastCGI is serving myself, much less kill them.  So I'm stuck while my sites are sometimes serving 500s and sometimes not.

I have simple questions, like what processes/pages are currently stuck?  How can I enable the mysql slow query log?  Why isn't the max_execution_time config variable being honored on CentOS?  I have asked 4 support reps these questions with no answers.  Only - I've put this on the ticket and escalated the priority.  These are questions I would love to troubleshoot, but they don't give me enough authority to do so or can't tell me how to exercise it.

I've started the process of putting everything behind CloudFlare, to help reduce traffic and hopefully displace some 500 errors.  I'm doing everything I can but am extraordinarily frustrated.  Time to call support a again...

[kjones75]

I've started the process of putting everything behind CloudFlare, to help reduce traffic and hopefully displace some 500 errors.  I'm doing everything I can but am extraordinarily frustrated.  Time to call support a again...

I've been through the same thing. I've tried blocking the IP addresses of suspicious activity (pretty much all of Russia and China must be blocked at this point), and I've installed Cloudflare (which didn't seem to have an effect at all, to be honest). Blocking IP addresses seemed to help temporarily, but the problem always returns.

I even moved 2 of my 3 websites that were sharing my LP account to another host thinking that would reduce the resource usage... those two websites are working great on my new host, but the one that remains on LP has gotten worse instead of better. The 508 errors used to be intermittent, now they are constant. I can't even log into my cPanel or FTP. And wouldn't you know it? I just renewed my LP hosting for another year a few days ago.

The server should not be having this much difficultly serving my small low traffic websites. It's not like I'm Facebook or something. My websites have never been super popular, and they've never used a ton of bandwidth or disk space. Ever since sites like Facebook and Twitter came along my sites have gradually lost more and more regular visitors. So I find it very difficult to believe my small websites are putting a strain on the server when they never did in their heyday several years ago.

I've been hosting the same 3 websites on my Lunar Pages account for almost 20 years now (since 2001). I've recommended LP to many people. But it looks like it's time for me to leave.

[MrPhil]

Has anyone on the support side been able to explain just what a "stuck process" or "stuck table" is? I presume something has gotten into an infinite loop. For a database, it would seem to be an error in the MySQL code (nothing you can normally do about it, unless you're doing some very advanced procedure stuff), but for a process, it might be something in your code. I suppose that it could also be some sort of deadlock over resource contention, but again, that's almost always an OS issue and not end-user code.

Has support offered any guidance on what might need updating in .htaccess or php.ini when your server is upgraded or account moved? Some obsolete settings may cause a 500 error, but I'm worried that from the sound of it, there are old settings that are simply ignored, or worse, do the wrong thing now. That should be the responsibility of the hosting company (LP) to at least document such things, if not scan customer accounts.

[virpus123]

The 500 Internal Server Error is a very general HTTP status code that means something has gone wrong on the website's server, but the server could not be more specific on what the exact problem is.

Are You the Webmaster? See Fixing 500 Internal Server Error Problems on Your Own Site towards the bottom of the page for some better advice if you're seeing the 500 Internal Server Error on one or more of your own pages.