Yes, AFAIK the default for memcached is 1024 connections. I graph connections using a munin plugin, and they have never been higher than 400. Additionally, when this has occured it has only effected one apache in the cluster at a time. I forgot to mention that the only way we have found to correct this is to restart apache. It does not appear to autocorrect. My best guess at what is happening, is that the connections are stale, but from the client, it believes that the connections are still active.
On Mon, Nov 3, 2008 at 11:18 PM, dormando <[EMAIL PROTECTED]> wrote: > > You sure you're not hitting the connection limit for memcached? > > you should do a little more troubleshooting to discover exactly what's > going on there... > > -Dormando > > On Mon, 3 Nov 2008, [EMAIL PROTECTED] wrote: > > > > > We are using memcache in a couple different ways, using the php client > > with mod_php. Once and a while it seems that something gets hinky in > > apache/mod_php/memcache client using persistent connections, and calls > > to $memcache->getServerStatus() will return false, which in our code > > causes an exception. > > > > I have a few questions: > > > > -Is there a way to configure the number of persistent connections to > > maintain? > > -Any other settings related to persistent connections? > > > > We have a small farm of 3-4 webservers currently going against a > > single memcached, and when I look for a connection count using lsof or > > netstat I get a number in the 255ish range -- is this a coincidence, > > or indicative of some built-in setting? > > > > I know there are a lot of variables here, our basic config: > > > > Centos5 > > httpd-2.0.52-41, running worker > > mod PHP 5.2.6 > > > > For now, I've disabled persistent connections, but I'm curious if this > > is going to significantly limit our scalability as time goes on. > > Recommendations or advice of others using memcache and php is > > appreciated. > > > > > > > > > > > > > > > > > > >
