Sudden large increase in process size, machine hang

2014-12-03 Thread Thomas Schulz
This last week we had a sudden large increase in the size of the named
process resulting in the machine running out of memory and hanging.
This is with bind 9.9.6 on a Solaris 10 Sparc machine.

I have posted in the past about a steady continuous growth in the
size of the named process. The Subjects for those posts are:
failed: out of memory
Process size versus cache size.
bind-9.10.0-P2 memory leak?  (I hijacked someone else's thread)

I have also filed a bug report about this continuous growth.

The following measurements show the steady continuous growth over
a little more than a month.

Thu Oct 16 14:01:42 EDT 2014 46014
Thu Oct 16 14:04:50 EDT 2014 99835
Thu Oct 23 20:06:01 EDT 2014317071
Thu Oct 30 09:31:31 EST 2014426647
Thu Nov  6 20:06:01 EST 2014532291
Thu Nov 13 11:00:46 EST 2014618143
Thu Nov 20 20:06:00 EST 2014701472

But then last week the following happened.

Wed Nov 26 08:06:00 EST 2014758702
Wed Nov 26 20:06:00 EST 2014   2099961
Thu Nov 27 08:06:01 EST 2014   3416580
Thu Nov 27 20:06:00 EST 2014   4618297
Fri Nov 28 08:06:00 EST 2014   5733163
Fri Nov 28 20:06:01 EST 2014   6801203
Sat Nov 29 08:06:52 EST 2014   No output from the ps command.

We were closed for thanksgiving when this happened. The only recursive
queries would have been for email processing by Sendmail. Monday morning
we had to reboot the machine by powering it down and back up.
The logs did not show any unusual messages from the named process.

Tom Schulz
Applied Dynamics Intl.
sch...@adi.com
___
Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe 
from this list

bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


Re: Sudden large increase in process size, machine hang

2014-12-03 Thread Warren Kumari
On Wed, Dec 3, 2014 at 1:39 PM, Thomas Schulz sch...@adi.com wrote:
 This last week we had a sudden large increase in the size of the named
 process resulting in the machine running out of memory and hanging.
 This is with bind 9.9.6 on a Solaris 10 Sparc machine.

This is probably going to be an annoying suggestion, but have you
considered using not a Sparc for this?

I was a huge Sun / Sparc fan (I *still* miss my Sun 4/360), but you
are in a definite minority using this for production BIND. Sometimes
there is safety in numbers...

W


 I have posted in the past about a steady continuous growth in the
 size of the named process. The Subjects for those posts are:
 failed: out of memory
 Process size versus cache size.
 bind-9.10.0-P2 memory leak?  (I hijacked someone else's thread)

 I have also filed a bug report about this continuous growth.

 The following measurements show the steady continuous growth over
 a little more than a month.

 Thu Oct 16 14:01:42 EDT 2014 46014
 Thu Oct 16 14:04:50 EDT 2014 99835
 Thu Oct 23 20:06:01 EDT 2014317071
 Thu Oct 30 09:31:31 EST 2014426647
 Thu Nov  6 20:06:01 EST 2014532291
 Thu Nov 13 11:00:46 EST 2014618143
 Thu Nov 20 20:06:00 EST 2014701472

 But then last week the following happened.

 Wed Nov 26 08:06:00 EST 2014758702
 Wed Nov 26 20:06:00 EST 2014   2099961
 Thu Nov 27 08:06:01 EST 2014   3416580
 Thu Nov 27 20:06:00 EST 2014   4618297
 Fri Nov 28 08:06:00 EST 2014   5733163
 Fri Nov 28 20:06:01 EST 2014   6801203
 Sat Nov 29 08:06:52 EST 2014   No output from the ps command.

 We were closed for thanksgiving when this happened. The only recursive
 queries would have been for email processing by Sendmail. Monday morning
 we had to reboot the machine by powering it down and back up.
 The logs did not show any unusual messages from the named process.

 Tom Schulz
 Applied Dynamics Intl.
 sch...@adi.com
 ___
 Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe 
 from this list

 bind-users mailing list
 bind-users@lists.isc.org
 https://lists.isc.org/mailman/listinfo/bind-users



-- 
I don't think the execution is relevant when it was obviously a bad
idea in the first place.
This is like putting rabid weasels in your pants, and later expressing
regret at having chosen those particular rabid weasels and that pair
of pants.
   ---maf
___
Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe 
from this list

bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users