old data don't expire on source server

[fredric]

Hi,
I have a primary source SP 8.1.21 server that replicate data(VMware) (stgrule replication) to a target server SP 8.1.21
I have different copygroup policies on the target server. On target the data are expired according the policy (30days), but on the source server the data aren't expired. It should be expired after 80days, but I can see data from September 2023 (show versions) for the VM I investigate. I have ran expired inventory on both servers, on the source server I can see that it only examine 150 objects and skip them all, nothing is deleted (every time/day it's run) for the filespaces belonging to the specific VM. I have tested to replicated the data all over with the forcereconsile=full, but no success. I have/had multiple cases with IBM support, but they can't find why it's not expired.
Have I missed something? I store the VMCTL in a different mgmt class/stgpool, but is also 80days. All other mgmt classes are max 80days in the same domain.
Any ideas are appreciated.

//Fredric

 

[fredric]

Hi,
I have a primary source SP 8.1.21 server that replicate data(VMware) (stgrule replication) to a target server SP 8.1.21
I have different copygroup policies on the target server. On target the data are expired according the policy (30days), but on the source server the data aren't expired. It should be expired after 80days, but I can see data from September 2023 (show versions) for the VM I investigate. I have ran expired inventory on both servers, on the source server I can see that it only examine 150 objects and skip them all, nothing is deleted (every time/day it's run) for the filespaces belonging to the specific VM. I have tested to replicated the data all over with the forcereconsile=full, but no success. I have/had multiple cases with IBM support, but they can't find why it's not expired.
Have I missed something? I store the VMCTL in a different mgmt class/stgpool, but is also 80days. All other mgmt classes are max 80days in the same domain.
Any ideas are appreciated.

//Fredric

I found out if I disabled the replstate on the filespace on the source server and ran "expire inventory nodes=<node name>", then the old objects were deleted.

Don't forget to update the status(replstate) to enable afterward. I don't really know I must do a full sync (replicate with forcereconsile=full) or if the ok to continue with the existing replication.

Why it don't work before can't even IBM support answer me.