strange issue

[okieunix1957]

i did have my cluster finally working then after I added the node. The cluster decide is does not know the name of the cluster.
Then another node 08 get this issue:
# ls -la pve
ls: cannot access 'pve': Transport endpoint is not connected

It like I lost all /etc/pve directory
d????????? ? ? ? ? ? pve

The above is ls -la output of that. i see the transport endpoint issue before but clears up by itself.

Any idea how to fix this issue other than rebuild the node.

 

[vopima2188]

i did have my cluster finally working then after I added the node. The cluster decide is does not know the name of the cluster.
Then another node 08 get this issue:
# ls -la pve
ls: cannot access 'pve': Transport endpoint is not connected

It like I lost all /etc/pve directory
d????????? ? ? ? ? ? pve

The above is ls -la output of that. i see the transport endpoint issue before but clears up by itself.

Any idea how to fix this issue other than rebuild the node.

Hi buddy, did you get any possible way out? I tried a lot but facing kind of the same issue and got stuck at this point.

Regards,
V. Smith TellGameStop MyCSULB

 

[okieunix1957]

Hi buddy, did you get any possible way out? I tried a lot but facing kind of the same issue and got stuck at this point.

Regards,
V. Smith

It turns out that corosync was messed up and had ensure corosync was corrected and make sure the version
was right. keep in mind the /etc/pve directory is a virutal and part of the pmxcfs filesystem. It came back on it's
own after I stop pve-cluster and corosync, then pmxcfs -l the restart corosync and pve-cluster. It came back ok.

I think this is because pmxcfs -l after stop pve-cluster and corosync before you make any changes to the corosync file. Keep in mind it's my understand that /etc/pve/corosync.conf overwrites /etc/corosync/corosync.conf file.
You need to do this to all nodes. THen it should come back.