before we address this perhaps you could try something for me as we could really do with a missing piece of evidence…

Could you go to the forum admin > toolbox > error log and scroll down to the bottom and empty the log. Then run the forum’s front page (group view) once. Go back to the error log and see if there are any entries and if so – is there a specific database entry. If so could you paste that in a reply please.

I cleared the SP error log, went to the forum page and then checked both, SP error log and server error log. Both have no new entries (SP error log remains empty).

(I removed all of the dupes in your original post by the way).

It is a shame nothing showed up as I was hoping we might discover if there is a genuine SQL error. Although of course the answer to that is no – not that time!

What is happening is that the mySQL server ‘went away’ – became unresponsive – and that triggered a DB error which we then tried to write to the SP error log which itself needs a database entry to be made which failed and tried to write an entry to the SP error log etc.. etc…

SO it goes on and on until the server makes a re-appearance.

What we don’t know is why the server became unresponsive and if the original error it was trying to write has any importance.

The first thing to say is that if I look through my php error log now I will find some of these ‘gone away’ messages – some of them related to SP and some not. Some within WP core sql calls and some in other plugins. You are not unique in that regard!

The next thing is that we did find we could minimise them by some carefully adjusted MySQL settings but (a) I did not do them and (b) I am afraid I have no idea what they were.

But most importantly – I rather thought that we had fixed the irrational looping with the error log write attempts quite a while back so I have to ask – which version of SP are you using?

Thanks so much for the detailed feedback, I am going to bookmark that for future reference.

Concerning SP version, I am using Version 5.2.6, Release 9877.

FYI, I just checked both error logs again, still no new entries. I’ll keep watching though.

OK – looks like I need to look at that code again to see what has shifted as I know we stopped that loop at one point.
Did you notice the dates on all of these groups of entries? If they were al around the same time then it could have been that the server was playing up for some reason at that point.

Yes, they were all within roughly 60 seconds, about twice a day. Just checked my server error logs and the error now reappeared, starting with:

[28-Mar-2013 01:15:33] WordPress database error MySQL server has gone away for query SELECT  * FROM wp_sfforums WHERE forum_rss_private=0 made by require(‘wp-blog-header.php’), require_once(‘wp-includes/template-loader.php’), include(‘/themes/paperpunch_pro_MWA/no-sidebar.php’), the_content, apply_filters(‘the_content’), call_user_func_array, sp_render_forum, sp_process_template, sp_load_template, include(‘/sp-resources/forum-themes/byw/templates/spGroupView.php’), sp_load_template, include(‘/sp-resources/forum-themes/byw/templates/spFoot.php’), sp_AllRSSButton, spdb_table, spdb_select

That’s the first entry after I rotated my logs. Lasted for about 45 seconds creating several hundred entries for the initially mentioned error, then went away. No new entries yet. I am investigating some other plugins for potential issues to see if anything uses up too many resources.

I have opened a ticket on the looping of the error log write so i can look into that.

It is perhaps curious and just a coincidence that both of the initial queries you have pasted here – i.e., the first reporting that the server has gone away – have both come from the RSS feed…. At least I assume a coincidence.

Have you checked your database tables recently? Might be worth running a check on them for errors and just optimising them generally.