[ClusterLabs] FYI: clusterlabs.org planned outages

Hi all, We are in the process of changing the OS on the servers used to run the clusterlabs.org sites. There is an expected outage of all services from 4AM to 9AM UTC this Thursday. If problems arise, there may be more outages later Thursday and Friday. -- Ken Gaillot

Re: [ClusterLabs] [EXT] Re: Fast-failover on 2 nodes + qnetd: qdevice connenction disrupted.

Hi! On " First of all, there no fencing at all, it is off." Maybe the default configuration should involve a fencing agent that sends an SMS like this to all admins: "Hey, get out of the bed and drive to work: nodeX has to be reset to continue working. You get this message, because you didn't

Re: [ClusterLabs] [EXT] Fast-failover on 2 nodes + qnetd: qdevice connenction disrupted.

Hi! Just some personal comment: If an application isn't cluster-aware (has no provisions to run in a HA environment), you may improve its uptime using a cluster, but you cannot really make it "HA". Just consider the app needs manual intervention after it crashed... Kind regards, Ulrich From:

Re: [ClusterLabs] [EXT] Re: "pacemakerd: recover properly from Corosync crash" fix

Hi! I wonder: Shouldn’t node fencing step in? What do other nodes say about the situation? Regards, Ulrich From: Users On Behalf Of Klaus Wenninger Sent: Monday, April 22, 2024 11:06 AM To: NOLIBOS Christophe Cc: Cluster Labs - All topics related to open-source clustering welcomed Subject: