When I had this problem I experienced more or less the following series
of events - this was awhile ago so my apologies
for the lack of exact detail (i.e. which argument, which range of vals,
1) try initdb and it fails (for some other config error than the semctl
invalid argument error).
2) fix the orig problem but now it fails because of the semctl invalid
3) each time retry, one of the semctl args bumps up higher and higher
each time, and the error repeats itself, in fact
the problematic arg gets more and more out of range.
4) finally i decided to try rebooting, wondering if that would wipe
whatever memory the semctl was using to determine
the initial value for its args.
5) this worked - so for me, the solution was, if initdb fails for any
reason, correct the problem, but then reboot before
It would be nicer of course to know where semctl is storing this
information and be able to change it there, etc. etc. etc.
> We are getting this same exact error-- always this
> semctl(35, 16, SETVAL, 536) failed: Invalid argument
> on the creating template1 step of initdb.
> On my Windows XP machine I had 7.4.1 working fine last month (service
> If I install 7.4.2 I get this error. If I reinstall 7.4.1 (rerun make
> install) initdb works. If I reinstall 7.4.2 it fails again with that
> semctl error. I tried rebuilding 7.41 from scratch and installing,
> thinking maybe it was a configuration difference when I built it last
> month, but no, 7.4.1 initdb still works.
> My coworker, on the other hand (not running it as a service), gets
> this same semctl error even with 7.4.1.
> This is preventing us from switching from mySQL to Postgres. Can
> anyone please offer a clue how to get around this?
> Eric Schwarzenbach
In response to
pgsql-cygwin by date
|Next:||From: Mary O'Donovan||Date: 2004-04-01 07:53:58|
|Subject: Problems with Postgresql 7.4.1|
|Previous:||From: Patrick Samson||Date: 2004-03-31 06:54:22|
|Subject: Re: Backend doesn't catch the next command, after SIGUSR2|