On Thu, Nov 7, 2019 at 7:35 PM Robert Haas <robertmhaas@gmail.com> wrote: > > On Thu, Nov 7, 2019 at 1:15 AM Ashutosh Sharma <ashu.coek88@gmail.com> wrote: > > @Robert, Myself and Prabhat have tried running the test-cases that > > caused the checkpointer process to crash earlier multiple times but we > > are not able to reproduce it both with and without the patch. However, > > from the stack trace shared earlier by Prabhat, it is clear that the > > checkpointer process panicked due to fsync failure. But, there is no > > further data to know the exact reason for the fsync failure. From the > > code of checkpointer process (basically the function to process fsync > > requests) it is understood that, the checkpointer process can PANIC > > due to one of the following two reasons. > > Oh, I didn't realize this was a panic due to an fsync() failure when I > looked at the stack trace before. I think it's concerning that > fsync() failed on Prabhat's machine, and it would be interesting to > know why that happened, but I don't see how this patch could possibly > *cause* fsync() to fail, so I think we can say that whatever is > happening on his machine is unrelated to this patch -- and probably > also unrelated to PostgreSQL. >
That's right and that's exactly what I mentioned in my conclusion too.
In fact, I suspect this is PostgreSQL successfully protecting itself from an unsafe situation.
Does the host have thin-provisioned storage? lvmthin, thin-provisioned SAN, etc?
