On a new VPS server set up at Go Daddy the Bind keeps stopping. There are no error messages logged. linux cos6 fresh install. bind verify is good, virtuslmin verify is is good. server has one GB of RAM 40gb memory. CP and free shows ram 100 % used but this may be because it is a virtual shared cloud server. Reported errors on the system just failing bind. Any suggestions would be good.
Sorry meant no reported errors
Submitted by andreychek on Thu, 04/25/2013 - 13:31 Comment #2
Howdy -- if you're using an OpenVZ-based VPS, you're likely running into resource issues.
You can check that by looking at /proc/user_beancounters, and seeing if any of the rows in the "failcnt" field are non-zero.
If they are, that means your operating system request a resource, and was denied that resource. We commonly see that on OpenVZ systems, and that can cause services to unexpectedly die.
If wish to use OpenVZ, the only fix for the above problem is to reduce your resource usage for that particular resource type (such as freeing up RAM), or to request additional resources from your provider.
Thanks I'm out all the office that will check when I get back . I don't think I'm using all the ram even load it reports all used. There is a 9 gigs of storage used and 40 severs.bind stops every five minutes . Also Pro fTP and clam av stop.
Submitted by andreychek on Thu, 04/25/2013 - 13:54 Comment #4
Yeah, what you're describing is typically caused by using OpenVZ and running low on resources. Folks using other VPS types, and dedicated servers, don't typically see any problems with services crashing -- BIND, ProFTPd, and ClamAV are all solid daemons.
The disk space doesn't usually matter, though RAM does. With OpenVZ, not all RAM is considered guaranteed... so what matters is making sure you're never using RAM that isn't guaranteed, or else the services could crash.
I've upgraded the RAM to 2GB and the BIND has dtopped crashing. total used free shared buffers cached Mem: 2048 2048 0 0 0 0 -/+ buffers/cache: 2048 0 Swap: 0 0 0
Checked beancounter and all failctn are 0
I may have an additional problem with files restored from backup on another server that was failing and I have imported DNS info from the old server and now I have invalid NS records and DNS errors. If I delete all BIND records, can I rebuild all servers with the correct records?
Submitted by andreychek on Fri, 04/26/2013 - 14:36 Comment #6
I'm glad to hear that the additional RAM helped.
Regarding BIND -- if you disable the BIND DNS Domain feature in Edit Virtual Server -> Enabled Features, and then re-enable it, it will re-create your records.
If you have any other questions -- it looks like you're using Virtualmin GPL there. The Support area is for folks using Virtualmin Pro... if you're using Virtualmin GPL, we'd encourage you to use the Forums for support.
We monitor those, along with lots of wonderful folks in the community. Thanks!
Yes,it is GPL. I posted from my cell phone and thought I had posted in the forum. Sorry about that. Thanks for the help.