Thread: checkpoint hang in 7.2b4
I was running a stock 7.2b4 Postgres with pgbench. RedHat Linux 7.2, stock everything. Dual PIII 650MHZ SMP kernel 1G RAM EXT3 file system Adaptec SCSI disks, one for /base and ont for/pg_xlog pgbench scaled to 50 ./pgbench -n -t 100 -h $HOST -c 100 bench It hung in the checkpoint. I wasn't able to trace it.
mlw <markw@mohawksoft.com> writes: > ./pgbench -n -t 100 -h $HOST -c 100 bench > It hung in the checkpoint. I wasn't able to trace it. Hung, or just took longer than you expected? I don't see how you expect us to do anything with that report. I can assure you I've run quite a few pgbenches on stock RH Linux 7.2, with no sign of a problem. So you'll have to do some investigation of your own, if you see a problem... regards, tom lane
Tom Lane wrote: > > mlw <markw@mohawksoft.com> writes: > > ./pgbench -n -t 100 -h $HOST -c 100 bench > > > It hung in the checkpoint. I wasn't able to trace it. > > Hung, or just took longer than you expected? No, hung, 20 minutes. Load average dropped to zero. > > I don't see how you expect us to do anything with that report. > I can assure you I've run quite a few pgbenches on stock RH > Linux 7.2, with no sign of a problem. So you'll have to do some > investigation of your own, if you see a problem... Hey I didn't have the debugger going, I was doing a performance test. I will look into it if it happens again. This wasn't a request for help, it was a heads up. If what I saw was anomalous that's great, but if it isn't I want to make sure that you guys heard about it, just in case.
mlw wrote: > > I was running a stock 7.2b4 Postgres with pgbench. > > RedHat Linux 7.2, stock everything. > Dual PIII 650MHZ > SMP kernel > 1G RAM > EXT3 file system > > Adaptec SCSI disks, one for /base and ont for/pg_xlog > > pgbench scaled to 50 > > ./pgbench -n -t 100 -h $HOST -c 100 bench > > It hung in the checkpoint. I wasn't able to trace it. I had a similar experience once with -c 5. I sent output of "ps ax| grep post" to this list and Tom claimed it to be fixed in CVS head. ------------ Hannu
Hannu Krosing <hannu@tm.ee> writes: >> It hung in the checkpoint. I wasn't able to trace it. > I had a similar experience once with -c 5. I sent output of > "ps ax| grep post" to this list and Tom claimed it to be > fixed in CVS head. Oh, you're thinking of that three-way deadlock condition. I've forgotten, was that in b4? regards, tom lane
On Sat, 2002-01-12 at 21:21, Tom Lane wrote: > Hannu Krosing <hannu@tm.ee> writes: > >> It hung in the checkpoint. I wasn't able to trace it. > > > I had a similar experience once with -c 5. I sent output of > > "ps ax| grep post" to this list and Tom claimed it to be > > fixed in CVS head. > > Oh, you're thinking of that three-way deadlock condition. I've > forgotten, was that in b4? Yes. -------------- Hannu