| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | "Jawarilal, Manish" <Manish(dot)Jawarilal(at)dell(dot)com> |
| Cc: | "pgsql-bugs(at)lists(dot)postgresql(dot)org" <pgsql-bugs(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: pgbench bug / limitation |
| Date: | 2020-05-28 17:31:30 |
| Message-ID: | 7221.1590687090@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
"Jawarilal, Manish" <Manish(dot)Jawarilal(at)dell(dot)com> writes:
> OS: Windows 10
> C:\Program Files\PostgreSQL\12\bin>pgbench.exe -c 120 -r -T 3600 -h <DBServerName> -p 9432 -U DBUser testdb
> too many client connections for select()
Yeah ... pgbench can support fairly large connection counts on platforms
that have ppoll(), but Windows does not. Without ppoll() we fall back
to select() which may not allow more than 100 or so.
Having said that ... it looks like pgbench thinks it can override
Windows' default setting:
#ifdef WIN32
#define FD_SETSIZE 1024 /* must set before winsock2.h is included */
#endif
and some googling confirms that indeed that should work. How did you
build or come by this copy of pgbench?
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Joe Conway | 2020-05-28 17:50:15 | Re: BUG #16466: Valgrind detects an invalid read in dblink_open() with a cursor inside a transaction |
| Previous Message | Jawarilal, Manish | 2020-05-28 15:30:55 | pgbench bug / limitation |