Old issue with "server has gone away for query"?

[Rarst]

Yesterday host threw more memory at my account, didn't help. Another load spike this evening, just ended.

I remembered I also have SSH access (kind of), any good commands I can run to determine what is eating resources?

Been trying some from this post but plenty (like top) don't run on throw errors (mostly with sort).

[Rarst]

Yuck.

Anyway after two days of back and forth with support they not only failed to find a problem, they failed to find an issue.

I also googled up post form person that did get answer out of them in similar situation. Basically average of 300 sites per server and they consider 1k visitors a day too much.

Anyway I guess I bored everyone to death with my constant issues so I will go sulk for some time and then start a topic, looking for a decent host. :)

[Rarst]

First meaningful information from support in two months:

It is found out one of the shared user used up the Apache resource as they scheduled lots of scheduled article from WordPress.

Will see if this deep tautology solution chases issues away.

Anyway what can you even do with scheduled posts to overload 8-core server?..

[Jeffro]

I don't know, seems like that wouldn't be the source of too many issues. Have no idea why that would cause some page loads from not happening.

[Ipstenu]

That doesn't sound right for the reason why. I can accept 'Someone else on your shared host box is a dingus who's sucking up your CPU' (I've been that guy) but reading that other guy's post, it sounds like they're simply overselling their servers which is causing higher than acceptable load times, and instead of balancing, they're waiting for people to kvetch and THEN balancing. Which is ... bad. It's no unethical, it's just a bad idea and stupid and ... Move hosts.

[Otto]

The server's gone.
The server's gone away.

The server's gone baby,
The server's gone away.

You know you queried wrong baby,
And you'll be sorry someday.

The server's gone.
It's gone away from me.

The server's gone baby,
The server's gone away from me.

Although I'll still serve on,
Only static content I'll be.

The server's gone,
It's gone away for good.

Oh, the server's gone baby,
Baby, its gone away for good.

Someday I know I'll be over it all baby,
Just like I know the web should.

You know the RAM is free, free now baby,
It's free from your spell.

It's free, free now,
It's free from your spell.

And now that it's over,
All I can do is drop you to shell.

[andrea_r]

*raises lighter*

[Rarst]

@Ipstenu

Yep, move sounds like a good idea indeed. :) Hence topic on hosting and WordPress.

PS another (much smaller) slowdown overnight... issue got better but didn't go away.

[Rarst]

Two quick questions as I stubbornly continue to dig this issue. :)

Is over 200 Apache processes normal on shared server?

/usr/local/apache/bin/httpd -k start -DSSL

Is it normal for blog to spawn over 30 wp-cron.php instances? Not mine, server neighbors - so troubleshooting options are limited.

[Rarst]

If cron is indeed the issue here it seems there are some long-stuck tickets about it.

http://core.trac.wordpress.org/query...order=priority

I wonder if I can get someone managing those tickets in touch with my host.