High server load and dsq_sync_forum problem

The problem has existed for years with Disqus and cron. Here’s a beautifully disastrous example of what can happen and why: https://core.trac.www.ads-software.com/ticket/15148#comment:9.

I hope someone from the team can take a look soon before more servers blow up.

Here’s a relevant pull request in their git https://github.com/disqus/disqus-wordpress/pull/103.

I filed a new ticket with Disqus here explaining the problem. I’ve also pinged my Disqus contact. Hopefully, they’ll take it seriously and fix it soon https://github.com/disqus/disqus-wordpress/issues/113.

Solved with a single line. https://github.com/disqus/disqus-wordpress/pull/115

If you add a wp_clear_scheduled_hook( 'dsq_sync_forum' ); before adding the schedule again in disqus.php, the wp_options cron maintains only one element.

Rafa, it seems to be a well solution to this issue. Disqus should update their plugin quickly…

Thanks!

I’m still testing the sync after adding this cleanup. No new comments so far. I’ll keep you posted ??

The solution offered right above will just continue clearing and postponing and never running the Disqus sync cron job on a busy server.

I have posted a PR to https://github.com/disqus/disqus-wordpress/issues/113 that properly resolved the issue for me.

Thanks for this, archon810. In first place, this is what created a busy server for me. If you wait 10 minutes, then you have 160 new elements in the wp_options cron. Which is no good at all, and it’s happening in a server with 0.10 load. If your blog has been running for 2 years, you have more than 16 million elements in the cron, exhausting the server.

I tried your solution before, but it was also adding more elements to the cron, which is not the best idea. There must be a way to remove old crons.

Just wipe your cron if it’s beyond salvageable like that. Empty the entry in the options table. If you have too much traffic that writes the same entry back over and over, stop all the servers for a moment, erase the entry, and restart them. That was the only way for me to get out of the undeletable cron.

And that’s why I was making it delete itself. I don’t want to be deleting it all the time. Disqus itself should be doing it. I hope they update soon also.

Are you all are using w3 total cache object caching?

It’s a yes for me.

Try turn off it and watch cron events with plugin Crontrol. It may help.

I have the same problem without using DISQUS.
Any idea whatelse can cause this similar:
Query_time: 119.743425 Lock_time: 3.012461 Rows_sent: 9 Rows_examined: 16759
SET timestamp=1478025975