Discussion created by ca.portal.admin on Jul 21, 2006
Hey Dennis!

Your second message came in as I was finishing this reply to your first
one; as you see below we fit your possible-hardware-issue scenario, with
one possible caveat.

We HAVE found the records in the journal, so IDMS DID at some point
think he put them to disk.

We then looked at the timestamp in the jrnl vs. the timings in TDMF and
they DID NOT line up (and yeah, we calculated and recalculated the GMT
offset in the jrnl vs. the local time in TDMF and still couldn't get
things to line up). The jrnl data happened a couple hours BEFORE the
volume was moved.

We then went and pfix printed the page from the source volume (still
sitting there unmolested), and the records WERE NOT on the source
volume. Means they either never made it there, or ""disappeared"" before
TDMF came along. TDMF appears to be off the hook for this one (with the
possible question of does he through some odd interaction mess up
Shark's caching processes).

We're still pondering, after having thrown out theories surrounding page
# overlaps (we use page groups), wormholes in space, and alien
abduction. Tagging this to a piece of hardware (would be Shark at this
point) is going to be messy.

Oh; one key thing I haven't mentioned yet; we're having unrelated
ABRT1125s in another application in this CV. CA has evidently never had
this ABRT reported, and it hints at buffer/bme corruption. Even though
it's associated with another app/different areas it's now our prime
suspect. Trying to reproduce to get a system snap for CA.