Ideally, i would like to have reproducible test at my hands to analyze it.
It is possible for you to provide me with it ?
As for debug dumps you have - they can't help as produced *after* deadlock
condition was found and cleaned by Lock Manager.
Thanks,
Vlad
No ths is not related to subject, this is another issue.
I have few such reports from customers but without test case to reproduce it.
I suspect this is related to garbage collection of a very long versions chain
(ten's backversions for the same record).
gstat -r could confirm it.
Regards,
Also, possibly related to this (appears on same production servers) - sometimes
all fb_inet_server.exe processes stop any disk i/o activity and wait for one
process, which eats 100% of one CPU core.
Stack trace for hung thread look like this:
fb_inet_server!down_grade+0x145
