> install seq 0 aru 0 high seq received 0 Oct 29 22:11:05 > flag0 retrans queue empty 1 count 1, aru 0 Oct 29 22:11:05 corosync > 29 22:11:05 corosync token retrans flag is 0 my set retrans > 22:11:05 corosync install seq 0 aru 0 high seq received 0 > retrans flag0 retrans queue empty 1 count 0, aru ffffffff Oct 29 > Oct 29 22:11:05 corosync token retrans flag is 0 my set > 1b received flag 1 Oct 29 22:11:05 corosync Did not need to originate any messages in recovery. > Oct 29 22:11:05 corosync previous ring seq 333572 rep > 29 22:11:05 corosync entering COMMIT state. > 22:11:05 corosync Storing new sequence id for ring 51708 Oct > corosync Saving state aru 1b high seq received 1b Oct 29 > Oct 29 22:11:05 corosync entering GATHER state from 0. > Oct 29 22:11:03 corosync entering GATHER state from 10. > 29 22:11:03 corosync entering GATHER state from 11. > Oct 29 22:11:02 corosync waiting_trans_ack changed to 0 Oct > (corosync cluster closed process group service v1.01) Oct 29 22:11:02 corosync Completed service synchronization, ready to provide service. > Oct 29 22:11:02 corosync Committing synchronization for > In corosync log I see repeated set of these messages, hoping these will give some more pointers. > I also ran tests with firewall off too on both the participating > I would recommend to disable firewall completely (for testing) and if everything will work, you just need to adjust firewall. > Verfiifed my config files nf and cib.xml and both have same
Also keep in mind that ports 5404 & 5405 may not be enough for udpu, because udpu uses one socket per remote node for sending. firewall is asymmetrically blocked (so node can send but not receive). config file on nodes are not in sync and one node may contain more node entries then other nodes (this may be also the case if you have two > clusters and one cluster contains entry of one node for other cluster) 3. > Keep in mind that if they are not directly connected, switch can throw packets because of MTU. And I do see my interface(eth1) on the box does have MTU as 1500 too.
> I am running with default mtu settings which is 1500. mtu is too high and fragmented packets are not enabled (take a > Subject: Re: daemon cpg_join error retrying > From: linux-cluster-bounces at > On Behalf Of Jan Friesse > How should I get access to Z stream? Is there a specific dir I should pick this z stream from?įor RHEL you are subscribed to RHN, so you should get it automatically, for CentOS, you should get it automatically. > I am running on Corosync 1.4.1 and pacemaker version is 1.1.8-7.el6Īre you running package version (like RHEL/CentOS) or did you compiled package by yourself? If package version, can you please send exact version (like 1.4.1-17.1)? What exact version of corosync are you running? Do you have latest Z stream? Subject: Re: daemon cpg_join error retrying The exact version is 'corosync-1.4.1-15'įrom: linux-cluster-bounces at On Behalf Of Jan Friesse I am running on the packaged version from RHEL 6.4. Next message (by thread): GFS2: fsid=M圜luster:gfs.1: fatal: invalid metadata block.Previous message (by thread): daemon cpg_join error retrying.daemon cpg_join error retrying Lax Kota (lkota) lkota at
0 kommentar(er)
