Hello,
I'm running Proxmox VE 3.3 ( pve-manager/3.3-1/a06c9f73 (running kernel: 2.6.32-32-pve) ) together with the latest FreeNAS as iSCSI backend in a two node cluster. After configuration and setup everything seemed to work fine until both nodes began to connect to the iSCSI backend every 20 seconds. At first I didn't really think about it as it's just a test environment, but it never stopped and so it's still running now for multiple days. I can see that the connection is clearly coming from the nodes to the backend and that the old connections are not dropped and data is transferred correctly. Everything seems to work except those connections every 20 seconds.
I wonder if I'm dealing with some kind of auto discovery here that went crazy? Or maybe some "alive checks" that are used to see if the backend is still alive.
Besides of that, everytime the backend is "probed" the iSCSI disk, which is used by the nodes, is reintroduced to the kernel. So I can see two messages in dmesg, that show the SCSI id and the size of the disk.
It looks like this:
And as said above it comes up every 20 seconds on both nodes...
I'm open to any idea as I already examined iscsid, the configuration and so on ...
KR,
G.
I'm running Proxmox VE 3.3 ( pve-manager/3.3-1/a06c9f73 (running kernel: 2.6.32-32-pve) ) together with the latest FreeNAS as iSCSI backend in a two node cluster. After configuration and setup everything seemed to work fine until both nodes began to connect to the iSCSI backend every 20 seconds. At first I didn't really think about it as it's just a test environment, but it never stopped and so it's still running now for multiple days. I can see that the connection is clearly coming from the nodes to the backend and that the old connections are not dropped and data is transferred correctly. Everything seems to work except those connections every 20 seconds.
I wonder if I'm dealing with some kind of auto discovery here that went crazy? Or maybe some "alive checks" that are used to see if the backend is still alive.
Besides of that, everytime the backend is "probed" the iSCSI disk, which is used by the nodes, is reintroduced to the kernel. So I can see two messages in dmesg, that show the SCSI id and the size of the disk.
It looks like this:
Code:
sd 8:0:0:0: [sdc] 536870912 4096-byte logical blocks: (2.19 TB/2.00 TiB)
sd 8:0:0:0: [sdc] 65536-byte physical blocksAnd as said above it comes up every 20 seconds on both nodes...
I'm open to any idea as I already examined iscsid, the configuration and so on ...
KR,
G.