Huge loads, runaway apache process Cpanel 6.0.0-S88

[ozzi4648]

Something is not right with Cpanel 6. I had to restart apache multiple times yesterday and already twice today. http processes are running wild to the point where sites are no longer accessible including by not limited to the loss of Squirrelmail and all GUI interfaces.

We have 35 sites on this account most of which are not even resolving. Apache is running WILD. Load averages are between 1.5 and 2.7 respectivily on a P4. On 35 sites?

Whats going on guys?

 

[ozzi4648]

[warn] [csd] filedescriptor (1024) larger than FD_SETSIZE (1024)

Can possibily be. There are only 35 sites on this box a few of which are running php and no cgi. RH 7.2. I have found quite a few people complaining about this issue.

 

[Sbee]

same problem here ?
at first we thought it was an account that had become popular , but it seemed it was cpanel 6 s88 , we also lost access to all client cpanels.
We now have r87 and the control panels are back , the load is better but not as low as they where before the upgrade.
I think cpanel need to get a dictionary and see the meaning of the word 'stable' .
These problems should not happen especially considering the price of a cpanel licence!

 

[xnull]

We only have about 45 accounts on the server right now, but have found that we can only run about 70 accounts max on our server. And we are paying a -lot- of money from our provider for what is supposed to be a fairly powerful server.. A lot of it to blame is Cpanel programs such as backups and logs.. I have problems like this all the time and have since we started the server 2 years ago nearly.

 

[shaun]

Have you checked the apache logs, have you checks the apache status in the WHM.


Sbee: Cpanel expensive? lol, Are you one of those 2 dollar hosts or somthing. Support, Updates, time and people cost money man. Come on....

 

[Sbee]

No
$60 -$70 month is expensive for buggy scripts that wipe out all users control panels on a stable version .
Maybe you think that`s ok ?
The problems only occurred when s88 was installed and have now gone again with r87.
Your right about support costing money, you tend to have increased support costs when all your customers lose their control panel .

 

[mesranet]

Since i upgrade to new Version 6 EDGE 77 and now to Stable 88, Latest Visitor and Webalizer are stop giving report and its effected to all user on my server. The report stop at 19th Feb. I already submit 2 ticket to [email protected] and also to RESELLER but until now no responce at all. I have to pay $99 per month to RESELLER but no support.

Is it FAIR enough ?

 

[Carl Ramirez]

netstat -np | grep SYN_RECV
netstat -np | grep FIN_WAIT
netstat -np | grep ESTABLISHED

check if thats going on?
how is the general status with http?

that should give an idea about the server load.

 

[Credanco]

Hm. We have 187 total accounts on our server. (don't think we're overloading it... the load average is around 0.33) It handles it fine without any problems. The last time our server went down was because of a kernel upgrade.

 

[ozzi4648]

Another restart of apache was necessary since i last posted the origional messages a few days ago. The damn apache processes are runnin wild httpd looping at 90% no sites accessible. I stop and then a start of apache is necessary to get things back into some sane condition. What the hell is going on with this? Im sick of this. Never had this problem before v6 came along!!!!

 

[Curious Too]

Originally posted by ozzi4648
Another restart of apache was necessary since i last posted the origional messages a few days ago. The damn apache processes are runnin wild httpd looping at 90% no sites accessible. I stop and then a start of apache is necessary to get things back into some sane condition. What the hell is going on with this? Im sick of this. Never had this problem before v6 came along!!!!

If it's a problem with V6 wouldn't it affect everyone who's running V6?

 

[ozzi4648]

Originally posted by Curious Too
If it's a problem with V6 wouldn't it affect everyone who's running V6?

It depends, i mean each server is different. We only have one server affected out of many. The others do not seem to suffer the same problem.

 

[bjarne]

Maybe you shuld start using the monitor tools available in WHM. Most likely there are some client running a script or something with a loop or something like that. In WHN you can use the top command and also trace a prossess and find what is it.

Once one of my servers suddenly got a high load and was on the edge of going down. I reloaded apache several times. But using th tool in WHM I found the sinner and suspended his account.

He was debuging a php script on the server, and there must have been a loop or something. THe klient removed the script and the server has been stable ever since.

 

[ozzi4648]

Originally posted by bjarne
Maybe you shuld start using the monitor tools available in WHM. Most likely there are some client running a script or something with a loop or something like that. In WHN you can use the top command and also trace a prossess and find what is it.

Once one of my servers suddenly got a high load and was on the edge of going down. I reloaded apache several times. But using th tool in WHM I found the sinner and suspended his account.

He was debuging a php script on the server, and there must have been a loop or something. THe klient removed the script and the server has been stable ever since.

Which monitor tool are you referring to?

Thanks

 

[rpmws]

Originally posted by ozzi4648
Another restart of apache was necessary since i last posted the origional messages a few days ago. The damn apache processes are runnin wild httpd looping at 90% no sites accessible. I stop and then a start of apache is necessary to get things back into some sane condition. What the hell is going on with this? Im sick of this. Never had this problem before v6 came along!!!!

Wel lI have to admit. I have been seeing this on 3 boxes almost one a day each now. Apache just hangs. It's not a mod_gzip loop or anything like that. My local monitor scripts don't catch it either. If it stays that way for 30 minutes the whole box goes down. It started at v6 but I can't be sure it has anything to do with it.

 

[ozzi4648]

Originally posted by rpmws
Wel lI have to admit. I have been seeing this on 3 boxes almost one a day each now. Apache just hangs. It's not a mod_gzip loop or anything like that. My local monitor scripts don't catch it either. If it stays that way for 30 minutes the whole box goes down. It started at v6 but I can't be sure it has anything to do with it.

Yep!

 

[rpmws]

WOW .. say what you really feel now ..

Like I said earlier I am not sure it's Cpanel related at all. It did start right after, in fact the same night we went to V6. It very well could be some apache exploit hitting us. I am not sure yet. I know that if it is Cpanel related I have not figured out what it's coming from.

I want to make it clear that my personal opinion, and I have been hosting since 97' and have used Alabanza, Plesk Ensim (still do) and Cpanel since 7-1-2000. My opinion still stands that cpanel is the best solution out there, even taking in consideration the problems we all know it has, support being one of them at times. But I have to admit that at times, quite a few over the years I have wanted to scream over cpanel related issues. I have considered moving away from cpanel but I feel lately things have gotten much better and when I say lately I mean last 2 years.

I know how you feel. You maybe are being a little hard on Annette. From what I know if her ..she can be pretty tough on a vendor, host ..or whatever ... "True*******" comes to mind. Regardless I want this crashing crap to stop!!!

 

[rpmws]

Originally posted by ozzi4648
Yep!

Well looks like you reconsidered and edited your post. Now mine looks stupid

 

[ozzi4648]

Just delete your message like i did! Good thing you didnt quote my message otherwise i would have had to kill you.