-
1. Re: URGENT: Need documentation for down_thread and up_thread
genman Dec 14, 2006 12:59 PM (in response to manojpn)This is a JGroups configuration detail
-
2. Re: URGENT: Need documentation for down_thread and up_thread
manojpn Dec 14, 2006 2:20 PM (in response to manojpn)Yes, it is a jgroups config and I am not sure what it does
-
3. Re: URGENT: Need documentation for down_thread and up_thread
belaban Dec 14, 2006 3:04 PM (in response to manojpn)This enables or disables the up or down thread for each protocol. By default, each protocol has a down and up queue, serviced by 1 thread each. To reduce context switching, these can be turned off individually, for each protocol separately.
I introduced this a long time ago, as I realized that the context switching overhead might be quite high.
In JGroups 2.5, the threadless stack will get rid of them anyway. -
4. Re: URGENT: Need documentation for down_thread and up_thread
manojpn Dec 14, 2006 3:38 PM (in response to manojpn)Thanks for your reply, really appreciate that. So are you suggesting that I should set these to false always for each protocol ? Is there any reason why I would set these to true for any protocol ?
I am using jgroups 2.2
Thanks! -
5. Re: URGENT: Need documentation for down_thread and up_thread
manojpn Dec 14, 2006 3:42 PM (in response to manojpn)Here is the current config I am using. We have 2 jboss clusters running on tomcat app server and three different machines. And every few days or so we get issue with cache being out of synch. Aur cache size seems to be around 1.2 MB in size. Do you have any suggestions for the config below ?
<UDP mcast_addr="224.0.0.100" mcast_port="48866" bind_addr="192.168.13.100" ip_ttl="64" ip_mcast="true" mcast_send_buf_size="150000" mcast_recv_buf_size="80000" ucast_send_buf_size="150000" ucast_recv_buf_size="80000" loopback="false" />
<PING timeout="2000" num_initial_members="3" up_thread="false" down_thread="false" />
<MERGE2 min_interval="10000" max_interval="20000" />
- <!-- <FD shun="true" up_thread="true" down_thread="true" />
-->
<FD_SOCK />
<VERIFY_SUSPECT timeout="1500" up_thread="false" down_thread="false" />
<pbcast.NAKACK gc_lag="50" retransmit_timeout="600,1200,2400,4800" max_xmit_size="8192" up_thread="false" down_thread="false" />
<UNICAST timeout="600,1200,2400" window_size="100" min_threshold="10" down_thread="false" />
<pbcast.STABLE desired_avg_gossip="20000" up_thread="false" down_thread="false" />
<FRAG frag_size="8192" down_thread="false" up_thread="false" />
<pbcast.GMS join_timeout="15000" join_retry_timeout="15000" shun="true" print_local_addr="true" />
<pbcast.STATE_TRANSFER up_thread="true" down_thread="true" /> -
6. Re: URGENT: Need documentation for down_thread and up_thread
belaban Dec 15, 2006 1:53 AM (in response to manojpn)"manojpn" wrote:
Thanks for your reply, really appreciate that. So are you suggesting that I should set these to false always for each protocol ? Is there any reason why I would set these to true for any protocol ?
I am using jgroups 2.2
Usually, yes, but with 2.2, I don't know. 2.2 was released in the fall of 2003, so I don't remember whether down/up_thread could be set to false. I suggest to upgrade to a newer version, e.g. 2.2.9.4, which is API compatible with 2.2 (even 2.3 and 2.4 are !). These newer stacks have all of their up and down threads set to false. -
7. Re: URGENT: Need documentation for down_thread and up_thread
belaban Dec 15, 2006 1:53 AM (in response to manojpn)Which cache ? JBossCache ?