Koozali.org: home of the SME Server

Obsolete Releases => SME Server 7.x => Topic started by: shaamone on February 01, 2007, 02:27:33 PM

Title: dnscache issue
Post by: shaamone on February 01, 2007, 02:27:33 PM

I'm having an intermittant problem with my internet connection dropping every ten minutes or so and have actually found it is happening at the same time as two dnscache processes appear on my SME 7.1 box.

I've read on these forums that this could be related to relays.ordb.org shutting down. I have found that actually sending an http request or ping to this host invokes the dnscache processes and eventually kills my adsl connection.

I'm guessing that the volume of requests generated by the dnscache processes is killing my router and initiating a reboot.

I have tried a few of the things mentioned ala removing relays.ordb.org from my setup, but this doesn't appear to do anything.. Has anyone come up with a workaround for this?

Title: dnscache issue
Post by: byte on February 01, 2007, 03:03:41 PM

I'm sure there was something in the Bug Tracker about this have a look over there, if it isn't raise a bug as it could be a potential bug. Thanks

Title: Re: dnscache issue
Post by: CharlieBrady on February 01, 2007, 03:22:49 PM

Quote from: "shaamone"

I'm having an intermittant problem with my internet connection dropping every ten minutes or so and have actually found it is happening at the same time as two dnscache processes appear on my SME 7.1 box.

There should always be two dnscache processes running on SME 7.1.

Title: dnscache issue
Post by: shaamone on February 01, 2007, 03:40:24 PM

Sorry, should have been clearer.. The two dnscache processes end up eating CPU useage and appear at the top of my process list in TOP.

Only way to resolve the problem is to killall dnscache.

Title: dnscache issue
Post by: CharlieBrady on February 02, 2007, 05:13:53 PM

Quote from: "shaamone"

Sorry, should have been clearer.. The two dnscache processes end up eating CPU useage and appear at the top of my process list in TOP.

Only way to resolve the problem is to killall dnscache.

No. All you need to do is stop looking up relays.ordb.org names. And/or apply the workaround given in the bug tracker.