Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 375 by [email protected]: Accept4 on FreeBSD 10 - sockets
stuck in CLOSE
http://code.google.com/p/memcached/issues/detail?id=375
We encountered that our memcached instances on FreeBSD 10 were rejecting
connections after being under heavy load.
We reproduced this behavior by writing a script that gets/sets random data
as well as rapidly opens and closes a lot of connections. The script will
reach the connection limit of our memcached instance in more or less
regular intervals.
While running this script we're monitoring netstat and memcached stats.
After running this script (sometimes for several hours) the server will
stop accepting connections and a high number of sockets (close to max
connections) is seen to be stuck in the CLOSE state. After closing the
memcached process these numbers (of course) disappear.
See the screenshot for the data we could gather.
Screenshot explained:
- The server has been started with "memcached -u nobody -c 8192 -m 512 -l
192.168.2.254 -p 11215 -vvv > & cache.log"
- Right window shows "tail -n 1000 cache.log"
- Top left shows "netstat -an | grep <state> | wc -l"
- Middle left shows the last "stats" command returned by the memcached
instance
- Bottom left shows errors returned by the test script
After looking through the memcached source we suspect the error to be in
the FreeBSD 10 kernel, but we would like to have a second opinion on that.
Attachments:
stuck.png 1.4 MB
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
---
You received this message because you are subscribed to the Google Groups "memcached" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
