I finally had the chance to upgrade our server to 2.1.2.13 last weekend.
Monday at noon I started the expiration-process.
I've been running the V3-server whithout proper expiration since about
mid november last year, so I expect the first run to expire a lot. And it
does..

I have a couple of questions concerning this..

The fix in 3.1.2.13 seems to bring us back to where we were at V2
concerning reclamation. Is this the same slow (but working) algorithm just
wrapped into V3?

Back in V2 the 'q proc', in addition to examined an deleted objects, also
reported which node was currently being processed.
This gave us a clue on how far the processing had gone.

In V3 this information is missing. Why? Is expiration no longer done
node by node?

This could be done by making sure that all nodes/filespaces had a file
with a no longer existing mgm-class. Running expiretion in "verbose" mode
would give the same clue from the act-log.. But that seems to be a silly
thing to do to get this information...

And, from V2.. I can't remember if the number of examined objects counts
up to the total number of objects in the db.... If it does, this could
also give a clue on how log the process will run.

With the next PTF.. Will we see a quick AND working expiration-algorithm?

Anyone?

                           Tom