I'm looking into an issue with failover not working during a recent outage on 1 of 2 CSRs used as DMVPN hub routers in Azure. The failover consists of event manager using the guestshell and invoking node_event.py scripts. This worked in the past, but failed for some reason recently (Mar 15, so I don't think it's related to the recent Azure auth issues) . While looking into the issue, I saw that the /dev/loop3 partition is 100% used, put over the edge by yum updates apparently. If I run a yum clean all it goes down to 92, and if I run an yum update it hits 100 again and the updates all fail. I'm not sure if this is in any way related to the failover issue or just another issue I should look at, but has anyone else had issues updating their guestshell or running out of space on a dev loop partition? The best I can tell, these aren't partitions I can resize. I don't know how their sizes are determined during the deployment. It looks like it's pointing to
[guestshell@guestshell events]$ sudo losetup /dev/loop3
/dev/loop3: []: (/bootflash/iox/repo/guestshell/extract_archive/guestshell.ext2)
thank you,
Bill
UPDATE - it looks like the space was being consumed by some Azure extensions, Network watcher and Linux diagnostic.
