(1) Do you have any of those third party process monitor scripts installed on your server?
(2) Does your server have anything installed which just restarts processes based on load/number etc?
(3) If you have any firewalls installed can you disable them for a few weeks and see if that sorts your issues?
(4) What do you see in apache error logs when this happens?
You need someone monitoring your server (and i am sure hostgator and others would never do as it's not really a managed server that they do in name of managed services) to get over this issue which from what i have noticed is a perennial problem with your install.
I know you wouldn't take this advice, but shift to PHPCow Managed Dedicated and give them a try. Just ignore server specs when it comes to pricing. There are companies doing celeron at price of dual cores.
First of all, it is a managed server.
Second, like every other host I've used....you guys always blame it on hosting.....and every single host either says you guys are crazy, or you have no clue what you are talking about. I'm not trying to be confrontational......but when We've switched hosts 4 times at the request of PHPCOW, and the issues crop up at every host.....and every host works with us.....but PHPCOW won't....its tough.......then we have you saying things like "We can fix it, we know whats wrong.....but we can't help you for free.....etc"....its really really frustrating, and it just really turns people off to PHPCOW.......I know I'm not alone in this, as i see a lot of people here with the same issues and frustrations.
If i haven't seen the issues spread across 40+ for over two years now I just can say what i have said and believe me it's not based on what my DC tells me for sure as every DC wou;d say it's Script and not HW/server setup.
Here's what a 300GB/Mo+ (current projection) Traffic server with one PHPCow site (Managed Server):
top - 10:51:07 up 129 days, 13:34, 1 user, load average: 1.00, 0.82, 0.81
Tasks: 118 total, 1 running, 117 sleeping, 0 stopped, 0 zombie
Cpu(s): 11.8%us, 0.3%sy, 0.0%ni, 87.7%id, 0.0%wa, 0.0%hi, 0.2%si, 0.0%st
Mem: 2010676k total, 1563868k used, 446808k free, 186644k buffers
Swap: 4192944k total, 3952k used, 4188992k free, 934204k cached
and here's the daily traffic details
01.06. 511.59 MB | 5,739 MB | 6,251 MB
02.06. 880.59 MB | 10,684 MB | 11,565 MB
03.06. 746.94 MB | 8,477 MB | 9,224 MB
04.06. 629.39 MB | 7,357 MB | 7,986 MB
05.06. 931.92 MB | 10,639 MB | 11,571 MB
06.06. 766.41 MB | 9,170 MB | 9,936 MB
07.06. 561.93 MB | 6,936 MB | 7,498 MB
08.06. 615.41 MB | 6,593 MB | 7,208 MB
09.06. 827.05 MB | 9,448 MB | 10,275 MB
10.06. 760.33 MB | 8,627 MB | 9,387 MB
11.06. 940.83 MB | 10,793 MB | 11,734 MB
12.06. 992.07 MB | 10,335 MB | 11,327 MB
13.06. 1,056 MB | 10,397 MB | 11,453 MB
14.06. 774.87 MB | 8,158 MB | 8,933 MB
15.06. 860.52 MB | 7,760 MB | 8,621 MB
16.06. 1,085 MB | 11,491 MB | 12,576 MB
17.06. 1,208 MB | 12,366 MB | 13,574 MB
18.06. 1,225 MB | 11,977 MB | 13,202 MB
19.06. 1,274 MB | 12,368 MB | 13,642 MB
20.06. 1,313 MB | 13,515 MB | 14,829 MB
21.06. 1,036 MB | 10,259 MB | 11,296 MB
and here's the monthly traffic growth:
month rx | tx | total
------------------------+---------------+---------------
Mar '08 13,694 MB | 157,816 MB | 171,511 MB
Apr '08 13,569 MB | 154,266 MB | 167,836 MB
May '08 17,997 MB | 202,830 MB | 220,827 MB
Jun '08 19,226 MB | 205,398 MB | 224,624 MB
------------------------+---------------+---------------
estimated 26,931 MB | 287,717 MB | 314,648 MB
You can check how long server has been up ... it never went down ever since it went live. That's what i call Managed Server in true sense though general perception of "Managed" deal is Hold Your Hand type of service.
My site has never gone down (the explosion and fire incident a few weeks ago notwithstanding)....i've never had a single issue with VBulletin.....only with PHPCOW.....now if we were having issues with both, that would be one thing......but its always just PHPCOW.
This new error we are getting is soooooo frustrating.....I've tried reinstalling the patch and that has no effect on the error at all.
Try to see error_logs whenever this happens and see if that holds any clue.