|
|
What determines ulimit settings on HP-UX?
What determintes ulimit settings on HP-UX?
For example, AIX has file /etc/security/limits.
Is there something similar on HP-UX?
We have two HP-UX servers, /etc/profile
are the same and do not contain references
to any external files, but output of "ulimit -a" is different:
server#1> ulimit -a
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 163840
stack(kbytes) 8192
memory(kbytes) unlimited
coredump(blocks) 4194303
server#2> ulimit -a
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 655360
stack(kbytes) 8192
memory(kbytes) unlimited
coredump(blocks) 4194303
Thanks
|
|
0
|
|
|
|
Reply
|
vafanassiev (40)
|
10/18/2005 1:41:43 AM |
|
Vsevolod Afanassiev wrote:
> What determintes ulimit settings on HP-UX?
> For example, AIX has file /etc/security/limits.
> Is there something similar on HP-UX?
For data and stack, the global limits are directly
derived from the maxdsiz/maxdsiz_64bit, maxssiz/maxssiz_64bit
kernel tunables. Processes get the global limits as their
hard limit if the parent did not modify (lower) the hard
limit. Soft limits start as the hard limit and can be modified
by processes as well. If the parent modified the hard limit,
that limit is propagated to the child (if possible, cases like
64-->32 exec don't always allow this, in which case the more
restrictive limit is applied (32-bit system limit vs. parent
limit).
If you're on 11i v1.6 or later, "man maxdsiz" or "man maxssiz".
If you're on 11i v1, take a look at:
http://docs.hp.com/en/TKP-90202/re34.html
Don
> We have two HP-UX servers, /etc/profile
> are the same and do not contain references
> to any external files, but output of "ulimit -a" is different:
>
> server#1> ulimit -a
> time(seconds) unlimited
> file(blocks) unlimited
> data(kbytes) 163840
> stack(kbytes) 8192
> memory(kbytes) unlimited
> coredump(blocks) 4194303
>
> server#2> ulimit -a
> time(seconds) unlimited
> file(blocks) unlimited
> data(kbytes) 655360
> stack(kbytes) 8192
> memory(kbytes) unlimited
> coredump(blocks) 4194303
>
> Thanks
>
|
|
0
|
|
|
|
Reply
|
Don
|
10/18/2005 2:38:18 PM
|
|
> For data and stack, the global limits are directly
> derived from the maxdsiz/maxdsiz_64bit, maxssiz/maxssiz_64bit
> kernel tunables. Processes get the global limits as their
> hard limit if the parent did not modify (lower) the hard
> limit. Soft limits start as the hard limit and can be modified
> by processes as well. If the parent modified the hard limit,
> that limit is propagated to the child (if possible, cases like
> 64-->32 exec don't always allow this, in which case the more
> restrictive limit is applied (32-bit system limit vs. parent
> limit).
>
> If you're on 11i v1.6 or later, "man maxdsiz" or "man maxssiz".
> If you're on 11i v1, take a look at:
> http://docs.hp.com/en/TKP-90202/re34.html
also see swapmem_on(5).
--vvp
|
|
0
|
|
|
|
Reply
|
Pai
|
10/19/2005 6:50:23 AM
|
|
|
2 Replies
1339 Views
(page loaded in 0.075 seconds)
|
|
|
|
|
|
|
|
|