Hi everyone,
Recently, we decided to upgrade proxmox (from 4.x to the latest 5.2). There are 2 nodes configured. One went down after rebooting, the other went fine. At that point, I was a bit surprised, same machines same manipulation, but anyway...
For the broken one, we had to rename the logical disk "shared/pool" to "shared/gfs" because the device /dev/shared/gfs didn't exist, while it was trying to mount it (hmmm, weird). There was a device named /dev/shared/pool instead. After that, the broken node booted fine. So far so good, except we had tons of errors in syslog. It was mainly due to connection and mount errors. Indeed, 3 shared storages had a small question mark in the GUI. Two glusterfs respectively named "shared" and "slowshared", and an iscsi named "iscsi-lun".
To reactivate both glusterfs storages, we tried to resync, which worked: gluster volume sync <node> all. The question mark disappeared for both storages, and mount was possible again.
Then, we tried to restart iscsid service on both nodes, which also worked. The question mark disappeared too for "isci-lun" storage. But, despite the fact that it is enable & active, it is not use. Indeed, we still have log errors about isci connections that are failing. This error is repeated every 2 seconds in syslog: "iscsid: conn 0 login rejected: target error (03/01)".
Right now, we don't have any clue how to solve this last (hope so, at least) problem. Can someone help us on this ? Thanks !
Recently, we decided to upgrade proxmox (from 4.x to the latest 5.2). There are 2 nodes configured. One went down after rebooting, the other went fine. At that point, I was a bit surprised, same machines same manipulation, but anyway...
For the broken one, we had to rename the logical disk "shared/pool" to "shared/gfs" because the device /dev/shared/gfs didn't exist, while it was trying to mount it (hmmm, weird). There was a device named /dev/shared/pool instead. After that, the broken node booted fine. So far so good, except we had tons of errors in syslog. It was mainly due to connection and mount errors. Indeed, 3 shared storages had a small question mark in the GUI. Two glusterfs respectively named "shared" and "slowshared", and an iscsi named "iscsi-lun".
To reactivate both glusterfs storages, we tried to resync, which worked: gluster volume sync <node> all. The question mark disappeared for both storages, and mount was possible again.
Then, we tried to restart iscsid service on both nodes, which also worked. The question mark disappeared too for "isci-lun" storage. But, despite the fact that it is enable & active, it is not use. Indeed, we still have log errors about isci connections that are failing. This error is repeated every 2 seconds in syslog: "iscsid: conn 0 login rejected: target error (03/01)".
Right now, we don't have any clue how to solve this last (hope so, at least) problem. Can someone help us on this ? Thanks !