Roger Zhou
Roger Zhou
Imaging the use case to add 10 more nodes to the existing 20 running nodes, it is not very good to restart them. Changed as `docker run --restart always`, and...
I have impression to suggest to move "./autogen.sh && ./configure" outside of the container execution environment. Only need in the host environment, let run-functional-tests prepares it once. It's a waste...
tw1:~ # rpm -q crmsh crmsh-4.4.0+20220208.5e6dca1d-1.1.noarch tw1:~ # crm --version crm 4.3.0
tw:~ # crm configure property maintenance=true WARNING: cib-bootstrap-options: unknown attribute 'maintenance' tw:~ # crm configure show | grep maintenance maintenance=true Instead of a WARNING, it's an ERROR. And, crmsh should...
Once I hit the error at the 2nd node when join, when the system disk is busy and has temporary access error in the test environment. I wonder the code...
Some improvement could have for `crm cluster crash_test` 1. more proper corosync2 split-brain simulation would be as below [[REF]] iptables -A INPUT ! -i lo -p udp -j DROP &&...
`crm node delete` should adopt most of functionalities of `crm cluster remove` as mentioned in PR890
Probably this is low priority. Not likely many user use this interface. > > `crm cluster remove tw3` works nicely so far. > > Well, I expect the following alternative...
For that, in crmsh code, `journalctl -o short-iso-precise` is needed instead of '-o short-iso'.
`crm cluster join` should have prerequisites, and not start in the very beginning if any cluster service is running, eg. sbd, corosync, pacemaker, corosync-qdevice
~ # crm cluster stop INFO: Cluster services stopped Basically, the term of "services" can be systemd services, ie, pacemaker.service, corosync.service, sbd.service. Also, it is widely used to refer to...