search

Jdesk / memcached

Automatically exported from code.google.com/p/memcached
0 stars 0 forks source link

Memcached 1.4.5 using 1.5G at boot time #142

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1.Start the memcached with the following command:
/usr/local/bin/memcached -u memcached -m 2048 -l 10.13.0.41 -p 11211

What is the expected output? What do you see instead?
1.Once started, if I launch:
ps waux |grep memcached |grep -v memcached
I get the following:
1018      8421  0.0  0.0 1527276  688 ?        Ssl  18:29   0:00 
/usr/local/bin/memcached -u memcached -d -m 2048 -l 10.13.0.41 -p 11211
The DRS is set to 1527276 just with memcached started. 
2.One hour later:
1018      8421  1.4 15.3 2800488 1274124 ?     Ssl  18:29   0:57 
/usr/local/bin/memcached -u memcached -m 2048 -l 10.13.0.41 -p 11211
The DRS is set to 2800488, higher than the 2048MB limit.

What version of the product are you using? On what operating system?
Memcached 1.4.5 (compiled at the same machine)
Redhat 5.4
libevent-1.1a-3.2.1
libevent-devel-1.1a-3.2.1

Please provide any additional information below.
I've got other linux systems with centos and redhat and when I launch 
memcached, the ps aux command looks like:
100       9178  0.7  0.4  95920 35972 ?        Ssl  18:40   0:00 
/usr/local/bin/memcached -u memcached -d -m 2048 -l 10.13.0.46 -p 11211
And the DRS is aprox 90MB, not 1500MB!

Original issue reported on code.google.com by ferranbo...@gmail.com on 2 Jun 2010 at 5:54

GoogleCodeExporter commented 9 years ago 
I've forgotten that 2 hours later I get a:
Segmentation fault

The last ps waux before segmentation foult was aprox:
1018      8421  1.4 15.3 3425778 1854523 ?     Ssl  18:29   1:45 
/usr/local/bin/memcached -u memcached -m 2048 -l 10.13.0.41 -p 11211

Original comment by ferranbo...@gmail.com on 2 Jun 2010 at 6:09

GoogleCodeExporter commented 9 years ago 
What version of redhat/linux are you running on the machines which don't have a 
problem?

Have you tried upgrading libevent on the RH5.4 machine?

Original comment by dorma...@rydia.net on 2 Jun 2010 at 8:51

GoogleCodeExporter commented 9 years ago 
The other machines where I've got memcached 1.4.5 running are:
CentOS release 5.4 (Final)
libevent-1.1a-3.2.1
libevent-devel-1.1a-3.2.1

Original comment by ferranbo...@gmail.com on 3 Jun 2010 at 7:07

GoogleCodeExporter commented 9 years ago 
Not much is expected to behave properly on libevent-1.1.  Can you upgrade 
libevent and try your bug again?

No reason to believe this is related, but even if we were to figure out what 
was wrong with this version and it 
*wasn't* libevent, you'd have bad luck running in production with this version 
of libevent.

Original comment by dsalli...@gmail.com on 3 Jun 2010 at 7:37

GoogleCodeExporter commented 9 years ago 
I've updated libevent package:
libevent-1.4.13-1
libevent-devel-1.4.13-1

I recompiled 1.4.5 version and now is working perfectly.

Original comment by ferranbo...@gmail.com on 4 Jun 2010 at 9:25

GoogleCodeExporter commented 9 years ago 
Good to hear.  We've talked about bundling libevent source to prevent this sort 
of thing.  It's manifesting in new 
and exciting ways.

Original comment by dsalli...@gmail.com on 4 Jun 2010 at 9:33

GoogleCodeExporter commented 9 years ago 
I just pushed the following changeset that cause memcached to refuse to start 
if I
detect a libevent 1.1 or 1.2..

http://github.com/trondn/memcached/commit/5ed9003ae236e6e2c2f3ad918794adb29607c3
05

commit 5ed9003ae236e6e2c2f3ad918794adb29607c305
Author: Trond Norbye <trond.norbye@gmail.com>
Date:   Fri Jun 4 12:01:16 2010 +0200

    Refuse to start if we detect libevent 1.[12]

Original comment by trond.no...@gmail.com on 4 Jun 2010 at 10:03