01-09-2001 09:07 PM
i integrated ob2 ver 3.5 with oracle 8.1.6 &
whenever i start the backup it gives this error Can not fork NETIO process ( Not enough space) => aborting.
i tuned the following kernel parameters :-
dbc_max_pct is doubled
shmmax is doubled
maxdsiz is doubled
swapspace is 1gb
enabled the oltp monolithic server
cell server is hpux 11 R-class with normal ethernet card connected to dlt4000 library
oracle clients are 2 nodes nclass cluster running ops 8.1.6 .
even i reduced the concurrency to 1 it failed after some time .
one more thing left if i start another backup
it gives lost connection .
i want to know is there any other kernel parameter to be tuned or i have to tune ob2 or
the network card can not handle the workload .
any help is higly appreciated .
Solved! Go to Solution.
01-10-2001 05:18 AM
the reason could also be that the proccess table is full. This is controlled by the kernel parameter nproc.
I would increase the kernel parameter maxuser that affects the nproc and other dependend parameters.
01-10-2001 06:02 AM
You mentioned you have doubled a couple parms; like dbc_max_pct. dbc_max and min_pct parms set up reserve buffer cache...I would be concerned that you may have reserved too much. Unless you have alot of memory (but you said you only have 256...)you don't want to set these too high...I'd recommend max/min of 10/5. If you've gone much over this..than your reserving buffer cache that other resources may not be able to get to. Also I'm wondering now about those other parms too. Although each system is different..and you know your system better than anyone.
Well those are my thoughts....
01-10-2001 08:50 AM
Is it true that you only have 256 Mb of Physical Memory?
If yes, I think that this is too low.
If you read the following url, this might help you:
Can you please provide the kernel parameters you've changed.
dbc_max_pct = ?
01-10-2001 09:01 AM
[Major] From: BSM@adhp3 "ora8AAP2adhp2" Time: 01/08/01 09:24:47 [61:1005] Got unexpected close from OB2BAR Backup DA on adhp2.
RMAN-10031: ORA-19624 occurred during call to
Do you have patch PHSS_22434 installed?
"[61:1005] Got unexpected close from INET on
To resolve your errors, add the following entries to your
# default: 10 minutes
Since this is an intermittent problem, either variable may resolve the problem.
01-10-2001 10:29 PM
However,nproc increased to 1024,dbc_min_pct is 5%,dbc_max_pct is 15%,tuned parameter set was applied ,& june 2000 general release patch was
installed. what makes me confused is that i
integrated ob2 on two nodes ,in the first node
it works fine but in the second it failes & oracle debug.log on the second node is attached for reference & in it you can see malloc failure , i increased shmseg to 64 & fired a backup i will be able to see it on
on Sat. 13/1/2000 .
01-12-2001 12:00 AM
Here you can find the latatest OBII patches.
Are the 2 nodes identical of sizes. Is the Oracle database the same size as the failing one?
Are the machines identical?
01-12-2001 08:43 PM
yes the 2 nodes identical in sizes ,it is an ops cluster ,i have 2 shared databses with 2 instances per node & the machines are 2
identical nclass-4000 servers .