Re: No space left to create named semaphores
Re: No space left to create named semaphores
- Subject: Re: No space left to create named semaphores
- From: Tron Thomas <email@hidden>
- Date: Tue, 21 Jun 2005 18:59:31 -0700
I'm not sure if any semaphores are leaked or not. I asked the developer
who works in the area that uses semaphores, and he said that most of the
semaphore created by his code should exist for pretty much the life of
the program. Other people may be creating semaphores that are not being
deleted. I have added some logging to see if any semaphores are being
deleted. So, far I haven't seen any evidence of semaphore deletion.
What tools and technique can I use to determine how many semaphores are
in use, and if there are any possible semaphore leaks?
John Francini wrote:
The semaphore space is controlled by the kern.sysv.sem* set of
sysctl(8) values.
The question, though, is how many semaphores is the program using, and
is it "leaking" semaphores? The default value appears to be 87381 in
Panther.
John Francini
On Jun 21, 2005, at 0:42, Tron Thomas wrote:
At work we have a program running on Mac OS X that routinely calls
sem_open to created named semaphores. After the program has run for
a while the sem_open command will fail with the result ENOSPC and the
corresponding error string "No space left on device".
What determines how much space is available for creating semaphores?
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden