On 2019-06-28 10:15, Tom Lane wrote:
tony@xxxxxxxxxxxxxxxxxxx writes:
I started an Analyze command on a database Wednesday evening at around
9:00PM. it is now Friday morning at 8:00 and it is still running.
...
I did try to execute:
SELECT pg_cancel_backend(4029);
and
SELECT pg_terminate_backend(4029);
but neither had any effect.
Hm, that's interesting. Can you get a stack trace from that process?
https://wiki.postgresql.org/wiki/Generating_a_stack_trace_of_a_PostgreSQL_backend
I am running PostgreSQL version 9.3 on Ubuntu 14.04 with 128GB of
memory, 800GB PCIe SSD for Database files, 1TB SATA SSD for WAL, 512GB
SATA SSD for system files.
9.3.what exactly?
(You do know that 9.3.x is out of support, so even if this
investigation
reveals a bug, we're not going to fix it in 9.3.x. I'm willing to look
anyway on the chance that there's a bug that also affects later
versions.)
regards, tom lane
Thanks for the offer to look at it Tom. Fortunately or unfortunately as
the case may be, after I installed everything to get the stack trace the
Analyze process actually finished. The max(last_analyze) did not change
and is still showing '6/27/2019 8:27 AM', so I am not sure what it was
doing all this time, but nothing seems the worse for it.
Thanks again.
