Discussion created by ca.portal.admin on Jul 20, 2006
Ummm. It's either Don Casey (APL) or David Brenner (ACS), but no
evidence of any journal oddness taking place.

The corruption is not easily explained by journal issues; two records on
the same physical page in different set occurances go missing, and the
rest of the set is fine. Seems a little too tied to the physical page
to be explained by something related to journal image processing (either
outbound or inbound).

More of the story. We have had two separate incidents of broken chains
in the past few weeks, and one common theme is both occurred during
periods when volumes were taking place. The most recent break was
isolated to a 6 hour window in which the specific volume the file sits
on was being moved, we have just located the STORE event in the
journals, and are attempting to tie that to when the volume was
physically moved.

The earlier breaks (at the beginning of the month) are also loosely
correlated to volume moves, but those journals rolled off before we got
past the fix-the-problem-so-production-can-continue stage and into the
what-the-heck-caused-this stage.

For completeness, and in case it triggers any thoughts out there, there
are some other commmon themes in these breaks:

Three sets of corruption seen: two out of the first volume move attempt
(early July), one out of the second (earlier this week).

Two applications (A and B) involved. Both A and B in first event, only
B in second. No other apps appear to be affected.

Same CV for all three apps (our other CV, equally active, has not had
any issues surface).

We're not sure what's coincidence and what's cause at this point.

Don Casey