ti 7. toukok. 2024 klo 14.30 Martin Grigorov (mgrigo...@apache.org) kirjoitti:
> Hi, > > On Tue, May 7, 2024 at 2:20 PM Martijn Dashorst < > martijn.dasho...@gmail.com> > wrote: > > > Hi! > > > > I was looking at our exception reporting in our application(s), and > noticed > > that when the CouldNotLockPageException is thrown, it only logs the stack > > elements of the request that keeps the lock. > > > > "... of the request that keeps the lock" > Isn't that the important stacktrace ? > I.e. it does not log the stacktrace of the current request but of the one > that keeps the lock, i.e. hangs. > +1 > > > > > > What I'd like to do is to construct an exception and fill that > exception's > > stack trace elements with the stack of the lock keeping request, and > > setting that as the cause of the CouldNotLockPageException. > > > > While this is a bit esoteric in nature, I think it is functionally > correct > > in that the page that keeps the lock is actually the cause of the CNLPE. > > > > This would make debugging the occurrences of the CNLPE a bit better, as > the > > exception itself is not informative, and you're actually interested in > the > > cause. So this would alleviate the need to peruse the logs of one's > > application, and you can just look in the reported exception itself. > > > > WDYT? > > > > Martijn > > >
