Slam Daniels

FiveTimes, when did you start wearing glasses??

Is it just me or is Bobhorse going to love this?

I have. And so have I.

It was a bright cold day in January, and the clocks were striking 12:30pm. Me and Dave were in a nuclear bunker underneath my house. I looked at Dave. He glanced at his wristwatch, saw that the time was 12:30pm and that the server load was increasing. He gave me the nod. I pulled the lever, delivering a surplus of 20gbit to the main server thruster. TLW lit up like a Dutch brothel in an autumn haze. Everybody continued to post shit about which Z-List celebrity they'd rather shag. And Dave saw that it was good. Dave U went home. Then Melinda Messenger came in and got out the squirty cream. (To be continued)

Only just! The server load shot up to 10.00 but it showed the 'frightfully sorry' message and then calmed back down. Looks like we're through the worst of it. I'll turn TLW Live back on once the server's been optimised and the server load is staying below 3.00. Cheers ears.

Hahaha! To me!

Courtesy of the nutters from b3ta.com.

So do I, it's crashing even when the server isn't busy. I've increased the max number of connections on MySQL and Apache and I've also decreased the KeepAlive time to 2. Things, at the moment, seem to be running swimmingly. Mind you, things have been running swimmingly about ten times today only to go to shite a bit later on. I think that it may be that Max_Connections for Apache, as it was only 100, was creating a bottleneck. Each instance of Apache had a KeeopAlive time of 15, so when the server was getting hammered, each instance was busy for 15 seconds. I'm waiting on the vBulletin dudes to suggest optimum MySQL and Apache settings so that we can draw a line under this. Server load averages: 1.87, 2.08, 2.09.

Back, just on the mobile. Progress seems to have been made, will update shortly.

I also noticed exim and spamd climbing the list a couple of times using up to 40% of CPU before dropping away again. This happens quite a bit.

Does anything look iffy?

We've got /etc/cron.daily/ ?

Not sure. There's no /etc/periodic/daily on this server apparently!

I've put that in place again. If the server load average is above 7.0 (which it's hitting now for some reason) it stops for a breather. Doing my head in now.

Had the same thought, but there's no crons set up on this server at all. EDIT: Obviously wrong, see later post. Also, scheduled task log in vBulletin shows: 4373 vBSEO Sitemap Generator 02:00, 9th Jan 2008 Google Sitemap [started] 4374 vBSEO Sitemap Generator 02:00, 9th Jan 2008 Google Sitemap Created 4372 Activation Reminder Email 00:03, 9th Jan 2008 Activation Reminder Emails sent. 4371 Daily Digest 00:02, 9th Jan 2008 Daily Digest Sent 4370 Birthdays 00:01, 9th Jan 2008 Birthday Emails Sent.

Yesterday's Apache Access graph. Spikes around this time also.

Server's been ok since 22:50pm and I've just found out, after spending two night and god knows how many hours trying to get the thing to work, that the answer was simple. Our host had installed CentOS with a kernel that vBulletin, for some reason, doesn't like. Upgraded it this evening and it looks like it's sorted. Really pissed off though, all that time wasted. I'm still taking the forum off in about ten minutes for, hopefully, 5-10 mins just to do some housekeeping on the database. I'm also going to tweak the server settings tomorrow so the server should be like shit off a shovel.

Sorry about the abrupt restart, I think I've pinpointed the problem with the server. Huzzah.

Seems to be steadying at the 1.70 mark now. Cheers DJLJ, smashy stuff.

Sorted, echo'd to it. Cheers DJLJ, we'll see how it goes.

It keeps reverting back to 32508. Any idea why it's doing that?

Could've been worse, I could've said notepad! Thanks

Am I ok to just do that using vi now do you reckon? Cheers for the help by the way.

32508