[Mongrel] mongrel goes out to lunch for about 5 - 10 minutes

Brian Doyle brianpdoyle at gmail.com
Mon Mar 5 15:30:30 EST 2007

Not sure if this is the list for posting questions like this but please let
me know.  We are using Mongrel 0.13.4, Mongrel Cluster 0.2.1  with Rails
1.2.2 with 3 different mongrel processes running in the cluster.  All
appears to be well for hours and then one of the mongrel processes appears
to hang and all incoming requests to that process timeout.  The requests are
not logged in the rails log, (production.log) and we do not see anything in
the mongrel.log so it is making it a bit hard to debug.  I know we can do
more debug logging for mongrel but we were attempting to wait to do that
until further investigation.

Eventually that mongrel process will "come back to life" and continue
serving requests again like nothing ever happened.  Nothing in any logs to
indicate the problem.  When I say eventually it appears anywhere from 5 to
10 minutes of time.  We are in the process of doing more debugging like
monitoring the network, database, etc. but just wondering if there is
something else we should be on the lookout for.   It appears that all
previous requests to that mongrel process are successful with no errors.
We are monitoring the 3 mongrel processes using Big Brother hitting the same
page.  This page doesn't require hits to the db, mail servers or memcache
(all services we use for other parts of the website.)  Thanks in advance.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/mongrel-users/attachments/20070305/455d9d3a/attachment.html 

More information about the Mongrel-users mailing list