|
I know that there is a mechanism inside of ASA that submits reports of the server crashing to Sybase (if we can remember where the switch is to flush the queue), but I'm curious where we would send reports of our application locking up if there are a large number of threads ( in this case ~15 threads ) that appear to be blocked inside of dbodbc12 ( two of which appear to be deadlocked against each other ). We obviously don't have symbols for dbodbc12 (and I haven't checked whether there's been a product update within the last couple months) but there's only so far that I can go when I see "dbodbc12 is waiting on EVENT HANDLE 0x1DD8" (with extremely limited ability to see our code lower down in the call stack). I know about the current "repro a bug" submission process that currently exists, although I'm not sure it is designed to submit minidumps (that may in the end still be caused by bugs within our program) |
|
You can use "dbsupport -sa" to submit all of the outstanding reports to SA development. To diagnose your client hang you will likely need to call Tech Support (see Contact Us link at the bottom of this page) if you want to submit the client minidumps since I don't think that the "repro a bug" process can take them. I think there has been at least one client hang issue fixed recently (within the past few months) so you should likely try a recent EBF first. Thank you for your fast reply. Something we might consider if we get a repeated lockup with a similar fingerprint, but for one-time race conditions it probably isn't worth the effort to open up a case. It would be nice if Sybae had a "public-symbol" server similar to what Microsoft does with symbols, might make it easier to at least trace through these kinds of crashes, but I suspect setting something like that up (and maintaining it) is a relatively monumental task.
(04 Jan '12, 15:49)
Erik Anderson
|
