How long is this going to take?!

[shwn]

I have the raw access logs. I'm hoping to see the error_log beyond just the past 300 messages. I just asked support for that, along with the memory limit increase. I have gotten a couple of those errors in the past few minutes for a 404 page, which is tiny, and I see no reason why anything going on concurrently makes that reasonable at all!

[Ron]

Unfortunately, unless something has changed in the last couple of weeks the error log is not maintained per domain beyond what you see. :-(

Look at the archived log for the time period when you were getting errors. Even if there are no errors logged in them, you should be able to see what "other" or simultaneous activity was going on.

You can also look at the very most current period in the top links.

Be sure that the archiving is on and not removing your logs monthly.

[shwn]

Nasir was kind enough to extract my domain's errors from the main file, as well as grant the limit increase. Thanks Nasir

[Ron]

Cool! How much error history were you able to get?

[shwn]

I just asked for Feb 13, the day I had the most problems accessing my site and lots of "Cannot allocate memory" errors, and that's what he sent.

Most errors that day were from one ip in France, user agent: Java/1.7.0
I see that quite often in my latest visitor stats.

[Ron]

That's awesome. The implication from support on the forums was that there was no historical error log info available... Specifically that CPanel wasn't generating splits, but from my recollection and impression it was implied that the data just wasn't available. This is good to know. Maybe something has changed.

Anyway, do you feel comfortable publishing that IP? I'll compare it with my spammer logs... or you could google it and see if it's listed as a bot, and see if it's listed as a good bot or a bad bot.

If it's a bot, do you have any controls in robots.txt? You could try to restrict its activities in some way. Some bots respect a crawl delay directive... Some ignore restrictions altogether.

Anyway, I hope you get this sorted out; I know it really bites when you've got an issue like this.

[shwn]

The words Nasir used:

extracted error logs from main error logs of Apache and wrote in the below file

so maybe not so easy...

It looks like the user with that ip first came to my site with Firefox for a page or two, then Java. The ip isn't listed as a bot.

Just now looking into controlling bots - hadn't felt a need until now.

Well, at least I've been able to access my site consistently. So hopefully the worst is over, whether it was more about the zone file problem or memory limits - who knows

[Ron]

It's possible that it's a browser pre-fetcher of some sort. Those things are always a pain in the arse. Some prefetch every darn link on the page as fast as possible.

Well anyway, good luck. I hope you've got it all sorted out now.