Re: slurpd memory leak (ITS#2620)

[quanah@stanford.edu]

--On Monday, June 30, 2003 10:28 AM -0700 Matthew Backes <lucca@csun.edu> 
wrote:

>> I've noticed that over the course of a day, slurpd's resident memory
>> size
>> increases.  It generally starts at 5MB, and by the end of the day
>> (when it is
>> stopped & then restarted) it is up to 10MB or more.  This would
>> indicate to me
>> that it has a (albeit slow) memory leak somewhere in the code.
>>
>> This happened tonight, after approximately 12,700 changes (MOD+ADD)
>> had been
>> performed during the course of the day.
>>
>
> Did you have a replica clause for a directory server that was down?  When
> it can't resolve changes as either complete or rejected, they sit in ram.
> Yes, that's probably still a bug.

No, all my servers were up, and I've had this happen a few times.  The fact 
that it doubles in size during the course of a day is rather disturbing.

--Quanah


--
Quanah Gibson-Mount
Senior Systems Administrator
ITSS/TSS/Computing Systems
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html