I would say best bet is to try out different values and benchmark in your app. We have a fairly busy cluster and run w/ 30 seconds w/ no problems at all. I am considering changing the default value (in my version) to 30 or 60 as I do think 5 is pretty aggressive. However, this is something that should be tuned in your app for best results.
<br><br>gw<br><br><div><span class="gmail_quote">On 9/22/07, <b class="gmail_sendername">Hanson Char</b> <<a href="mailto:email@example.com">firstname.lastname@example.org</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Hi,<br><br>I refer to the (KETAMA_HASH version) java_memcached-1.6. I realize<br>SockIOPool's the default maintenance thread sleep time "maintSleep" is<br>set to 5 seconds, which seems pretty busy. I am considering setting
<br>the sleep time to, say, 30 or 60 seconds but fear if this would cause<br>too many connections built up between the running of the maintenance<br>thread. The running application using the MemCachedClient is<br>processing from 30 to 90 requests per second.
<br><br>Any real life example on how safely this parameter can be set ? Caveat ?<br><br>Thanks in advance,<br>Hanson Char<br></blockquote></div><br><br clear="all"><br>-- <br>Greg Whalin