ca.portal.admin

Re: IDMS @ TDMF

Discussion created by ca.portal.admin on Jul 20, 2006
Steve:

Thanks for the info, this is comforting (but we still have our little
corruption problem).

One question; we are assuming our problem is related to heavy update
activity going on while the volume transfer is taking place. The
specific file in question can (per Perfmon) be averaging 5-10 writes
per
second over a 30 minute period; I have no idea what the instantaneous
peaks might look like. Does your system have files subject to that
level up write activity, or is your workload primarily retrieval?

Thanks;

* DQC *

Don Casey
APL Limited (David's customer).

P.S. For the amusement of the listening audience; we've seen incidents
where the broken chains are best explained by a ""write that didn't
happen""; e.g.; two set occurrences pointing (both next and prior) into
a
page where the two records in question just aren't there. Either
IDMS,
MVS, the Shark DASD, the EMC DASD, or TDMF dropped the ball somewhere.
Our DBAs are running some DBANs trying to get a better picture of what
happened when. Will advise the list if we ever reach an understanding
of what happened.

Outcomes