anyone have some insight?

Ray tech13 at powerweb.net
Thu Feb 6 20:40:25 UTC 2003


we have 2 dns servers here, both in the same room, and some software that=
=20
does a lookup on the software every 3 minutes and attempts to email/page=20
someone if any system is down for 'too long'.

it also show how long it look to get a look up for our main webserver

anyways, in the reports, someone noticed both dns servers response time w=
ent=20
from 5-50ms to 100-400ms.  we recently upgraded hardware&software on dns2=
 (a=20
few days before this), but haven't touched the hardware/software on dns1.
none of the other servers that this software watches show a similar oddit=
y.

doing a tcpdump on dns2 doesn't really show anything unusual, but then i=20
don't normally watch it on a dns server.=20

snip from dns1 status report
1/31/2003_4:48:03_AM 37652.2000387731 dns1.powerweb.net dns DNS 53 OK 19m=
s =20
1/31/2003_4:50:03_AM 37652.2014318634 dns1.powerweb.net dns DNS 53 OK 18m=
s =20
1/31/2003_4:52:04_AM 37652.2028256597 dns1.powerweb.net dns DNS 53 OK 17m=
s =20
1/31/2003_4:54:04_AM 37652.2042195602 dns1.powerweb.net dns DNS 53 OK 225=
ms =20
1/31/2003_4:56:05_AM 37652.2056139236 dns1.powerweb.net dns DNS 53 OK 250=
ms =20
1/31/2003_4:58:06_AM 37652.2070145486 dns1.powerweb.net dns DNS 53 OK 327=
ms =20

snip from dns2 status report
1/31/2003_4:48:03_AM 37652.2000392361 dns2.powerweb.net dns DNS 53 OK 6ms=
 =20
1/31/2003_4:50:03_AM 37652.2014324421 dns2.powerweb.net dns DNS 53 OK 7ms=
 =20
1/31/2003_4:52:04_AM 37652.2028261227 dns2.powerweb.net dns DNS 53 OK 6ms=
 =20
1/31/2003_4:54:04_AM 37652.2042209491 dns2.powerweb.net dns DNS 53 OK 346=
ms =20
1/31/2003_4:56:04_AM 37652.2056126505 dns2.powerweb.net dns DNS 53 OK 137=
ms =20
1/31/2003_4:58:05_AM 37652.2070132755 dns2.powerweb.net dns DNS 53 OK 216=
ms =20

and as far as i know, no one was in the office before 5am that day (we op=
en=20
at 8:30am).

its not technically broken, its not 'too slow' but i guess it seems odd a=
nd=20
possibly a sight that something is about to break.
cpu isn't pegged (its sometimes up to 3%) ram is mostly buffer, swap in=20
hardly used (5MB right now) and the harddrive is nearly not being used.
anyone have any ideas on what to check or look at to see what might be go=
ing=20
on?


More information about the bind-users mailing list