ASE 12.5 cannot start with more than 4.1 GB memory
This is ASE 12.5.0.2 on HP Tru64 (former Digital UNIX). We have 8 GB physical memory and 16 GB swap space. shm_max was configured as 6GB and now it is 7GB. But it didn't make any difference and ASE doesn't start with more than 2176000 (2K) page memory.
OS version is OSF1 V5.1 1885 alpha
Error message in errorlog is: os_create_region: shmat(8196): Not enough space
OS version is OSF1 V5.1 1885 alpha
Error message in errorlog is: os_create_region: shmat(8196): Not enough space
ASKER
Hi Bret,
It became more confusing after trying something else. Sybase starts with 7 GB memory when default data cache size is relatively small like 3GB. But let's say it doesn't start with 5 GB max mem and 3.5 GB default data cache giving the same message like shm_max is not big enough.
It became more confusing after trying something else. Sybase starts with 7 GB memory when default data cache size is relatively small like 3GB. But let's say it doesn't start with 5 GB max mem and 3.5 GB default data cache giving the same message like shm_max is not big enough.
ASKER
Hi,
Ignore my previous comment, because I forgot that ASE 12.5 behaves in a different way by default. The actual memory allocated is still less than 4GB (not 4.1 GB) and that system has 7516192768 bytes of shm_max configured in /etc/sysconfigtab (and also according to /sbin/sysconfig output).
Ignore my previous comment, because I forgot that ASE 12.5 behaves in a different way by default. The actual memory allocated is still less than 4GB (not 4.1 GB) and that system has 7516192768 bytes of shm_max configured in /etc/sysconfigtab (and also according to /sbin/sysconfig output).
ASKER
Hi,
HP tech support suggested to apply a patch and add two new parameters in /etc/sysconfigtab:
proc:
per_proc_address_space = 0x40000000000
max_per_proc_address_space
Now it is fixed. Maybe I should have sent the question on UNIX section. I didn't realize that it was OS rather than Sybase.
HP tech support suggested to apply a patch and add two new parameters in /etc/sysconfigtab:
proc:
per_proc_address_space = 0x40000000000
max_per_proc_address_space
Now it is fixed. Maybe I should have sent the question on UNIX section. I didn't realize that it was OS rather than Sybase.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Basic question, I know, but has the OS been rebooted since
it was changed?
-bret