Linux – A lot of TIME_WAIT connections? – CentOS

linuxPHPtcp

I have 4 web servers running php which saves sessions to a memcache over my private network. Today just about every hour I have my logs filled up with this error

ERROR - 2010-11-07 16:31:33 --> Severity: Notice  --> session_start() [<a href='function.session-start'>function.session-start</a>]: Server 10.xx.xx.xx (tcp 11211) failed with: Connection timed out (110) Session.php 212

I checked the memcache server and it doesn't have a lot of connections to it just a about 5 or so established connections, but when I check web3 server for an example it has 13446 connections in TIME_WAIT and about 6962 of those are port 11211. My question is, is this high number of time_waits causes my server to connection time out to the memcache server or is there something else I need to be looking at?

Best Answer

TIME_WAIT can be a side effect of NATing or of firewall configuration, but in neither case does it generally represent an issue, although it may be a symptom of a bigger problem. If you do in fact have a firewall separating the two machines, consider eliminating it.

Related Solutions

Linux – Cannot mount a CIFS network share on Ubuntu

Solved the problem.

It turned out I had an entry in /etc/hosts that made my DC hostname resolve to a different IP address than the primary reverse resolver.

IE:

192.168.1.160 server

While 192.168.1.160 did not resolve back to this host.

I edited the hosts entry to reflect the correct IP-adress, which solved the issue.

Linux – Network error with 65k of TIME_WAIT connections

FIN_WAIT (timeout for the FIN request acknowledgement) is not the same as TIME_WAIT (time to ensure that the socket is really not used anymore). And yes, with 65k ports in TIME_WAIT state, you only will be running out of TCP ports if you are using a single requester IP - as might be the case if all your clients are behind a NAT device. You also might be running of resources due to an overly populated transmission control block table - see this excellent even if somewhat dated paper for possible performance implications.

If you are really concerned about your sockets in TIME_WAIT state and do not have stateful firewalls between your clients and your server, you might consider setting /proc/sys/net/ipv4/tcp_tw_recycle, but you would sacrifice RFC compliance and might have interesting side-effects in the future due to that matter.

Best Answer

Related Solutions

Linux – Cannot mount a CIFS network share on Ubuntu

Linux – Network error with 65k of TIME_WAIT connections

Related Topic