So, it turns out the additional debugging we setup in that sssd package on the ipsilon servers isn't sufficent to show what the problem is. ;(
sssd developers would like us to crank up debugging to try and get logs of whats happening. Unfortunately, this will use a ton of disk space, which these vm's do not have much of. ;(
I was going to ask to grow the disk on ipsilon02, but I think perhaps cleaner would be to create a NFS volume and mount it on /var/log/sssd/ on ipsilon02. This way we don't need to actually take the machine down and it's really easy to back out.
So, can I get +1s to:
* make a NFS volume and mount it on /var/log/sssd on ipsilon02 * adjust ipsilon02's sssd to enable crazy levels of debug. * restart sssd and wait for the next failure. * profit!
+1s? other ideas?
kevin