removing domain dns zone crashs server (new bug?)

[zaitsev]

hi, i am trying to remove a parked domain but every time i try it, it crash's the server, this is what happened:

i have 2 domains, example.com and example.info, example.info is parked on example.com, i now want the .info to be on a different account so i login to cpanel and remove it from the account, the server freezes and the load shoots up to 60+ and ram usage 90%+

server reboots itself, i try removing it again from whm and the server crashs again, after it has settled down i i login to whm again but this time it example.info has gone from the parked domain list (so i think everything is ok now) and i then go to create a new account for the example.info but it fails and tells me that a dns zone already exisists for this domain, so i go to the dns zones in whm and there it is example.info, so i go to the delete zone menu in whm and try to delete example.info but it crash's the server, i have tried a few times and its still there, all it does is crash the server

has this happened to anyone before? is there a fix? or anyway i can get example.info moved to its own account?

thanks for your help and time
steve

[chirpy]

That's pretty odd. Have you tried checking in the error log /usr/local/cpanel/logs/error_log to see if there're any entries that indicate where the problem might be before the crash. I can only assume that it's getting into an infinite loop somewhere. If you're on the latest version of the cPanel tree that you use you might want to log a support ticket with cPanel or your license provider to have them investigate as it obviously should not be crashing your server!

[cpanelben]

What build are you using?

Have you opened a support ticket with us yet? If not then please do so ASAP.

Thanks.

[Mike2Own]

wow weird

i dont know if this is related or not but we have had a server crash several times over the last few days and i was digging through /var/log/messages and the very last thing the server did before the crash seems to be dns/cpanel related (updateuserdomain)

here is what i found in /var/log/messages

Aug 12 14:42:51 srv1 kernel: Unable to handle kernel NULL pointer dereference at virtual address 0000000c
Aug 12 14:42:51 srv1 kernel: printing eip:
Aug 12 14:42:51 srv1 kernel: c01ad696
Aug 12 14:42:51 srv1 kernel: *pde = 1f4ef001
Aug 12 14:42:51 srv1 kernel: *pte = 00000000
Aug 12 14:42:51 srv1 kernel: Oops: 0000 [#1]
Aug 12 14:42:51 srv1 kernel: SMP
Aug 12 14:42:51 srv1 kernel: Modules linked in: md5 ipv6 ipt_TOS iptable_mangle ip_conntrack_ftp ip_conntrack_irc ipt_REJECT ipt_LOG ipt_limit ipt_multiport
ipt_state ip_conntrack e1000 iptable_filter ip_tables usbcore ext3 jbd
Aug 12 14:42:51 srv1 kernel: CPU: 3
Aug 12 14:42:51 srv1 kernel: EIP: 0060:[<c01ad696>] Not tainted VLI
Aug 12 14:42:51 srv1 kernel: EFLAGS: 00010246 (2.6.10)
Aug 12 14:42:51 srv1 kernel: EIP is at __rb_erase_color+0x46/0x1c0
Aug 12 14:42:51 srv1 kernel: eax: f7764ca4 ebx: f7567388 ecx: 00000000 edx: f7764ca8
Aug 12 14:42:51 srv1 kernel: esi: dc166c9c edi: f7764ca4 ebp: f7764ca4 esp: cf8fff14
Aug 12 14:42:51 srv1 kernel: ds: 007b es: 007b ss: 0068
Aug 12 14:42:51 srv1 kernel: Process updateuserdomai (pid: 3726, threadinfo=cf8fe000 task=d3e7e020)
Aug 12 14:42:51 srv1 kernel: Stack: f7567388 f7764ca4 00000000 dc166c9c 00000000 c01ad85b 00000000 f7567388
Aug 12 14:42:51 srv1 kernel: f7764ca4 dc166d04 dc166c9c f7764ca0 e5744ce4 f885a359 dc166c9c f7764ca4
Aug 12 14:42:51 srv1 kernel: dc166d04 dbccf380 dc166d04 f885c770 dc166d04 dbccf380 f7f89200 c015ab52
Aug 12 14:42:51 srv1 kernel: Call Trace:
Aug 12 14:42:51 srv1 kernel: [<c01ad85b>] rb_erase+0x4b/0xf0
Aug 12 14:42:51 srv1 kernel: [<f885a359>] ext3_discard_reservation+0x59/0x70 [ext3]

[Mike2Own]

that server -^ is using WHM 10.1.0 cPanel 10.2.0-R82

[icontact]

This seems to be happening to a WHM that is installed on a VPS. Cant remove the DNS Zone at all and the server load just skyrockets.

Is there an update on what is causing this?

[checked]

Exactly the same happened to me last night. I was terminating one account and suddenly my server got hangged in the middle of the Termination and then I asked the NOC and they rebooted it. After that till now I am unable to rea reate that account because each time I try to create that account it gives me the following error :

Sorry, a DNS entry for xyz.com already exists, please delete it first (from all servers in the dns cluster)

I have checked the sever and didn't find any entry or file or DNS for this domain neither ant DNS Zone file nor it is in httpd.conf or anywhere else BUT Still it is not allowing me to recreate that account shows me the above error.

My WHM Version Details are :
WHM 10.1.0 cPanel 10.2.0-R82
RedHat Enterprise 3 i686 - WHM X v3.1.0

Any ideas ????

[NoDoze]

I'm having the same issue right now! OMG I'm at least happy to see I'm not the only one having this issue....

Any answers?

Is there a way I could manually delete it?

[checked]

I wonder if you also have the same Version of WHM ?

Also, I had contacted my NOC and suggested them to launch a Support Ticket at cPanel (because now cPanel only accepts Tickets from NOCs) and hope to get a positive reply soon

[NoDoze]

hehe...well...I am the NOC....sheeesh....

I have about 250 clients who are on that server, and are now recieving slow load times cause the cpu is overloaded...

[bullethost696]

I had a similar problem, it was a problem with my DC's resolver (Sagonet). Not to sure what they did but after my machine was rebooted everything worked again.

[chirpy]

A fix went into the RELEASE and upwards trees last night for a DNS looping issue. Have you tried upgrading?

[checked]

Noop........but will the upgrading issue will also solve my problme of re-creating the Web Account which I am unable to recreate. Each time I try to create it displays :

Sorry, a DNS entry for xyz.com already exists, please delete it first (from all servers in the dns cluster)

It seems like that cPanel's script also store the domains somewhere else (uncommon place) becasue as far as I am concerned I have removed the DNS entries and all other things where the acount reference was mentioned but stll no luck.

Any idea ??? where should look to resolve the issue.

[icontact]

I've upgraded to

WHM 10.6.0 cPanel 10.6.0-R16 which seems to have fixed the problem - WHM seems to be running better but the swap usage is currently at 98% - If its not one thing its the other.

Aaaargh.

Im running this release of WHM on a VPS

Current Memory Usage

total used free shared buffers cached
Mem: 8195448 7575804 619644 0 1027000 2180380
-/+ buffers/cache: 4368424 3827024
Swap: 2096440 2056228 40212
Total: 10291888 9632032 659856


Swap used to be under 1% before the upgrade to this new release though Im not sure at this stage if it is directly related.

Still quiet inexperienced at this so can someone shed some light on how I fix this high swap usage?

Thanks

[chirpy]

That's a serious problem, most likely a memory leak. Try running top and press c and them M and look at the top processes.

Common culprits:

httpd
named
cpanel services (though the upgrade should have sorted that out)

Another problem when you're memory thrashing like that, is that the original cause may be masked by the turmoil it has created. Restarting services down the list is sometimes the only way to restore order (or rebooting if you're into that sort of thing).