PCI compliance is a big job based on the number of rules it executes
However - how many job servers do you have and how many WorkItemThreads does each server have?
What MaxHeapSize do you have set?
Below is configuration:
BLManager Version = 8.1.02.233
Async integration = Enabled
Maximum concurrent jobs = 20
Number of jobs currently running = 0
Number of work item threads = 50
Available work item threads = 50
File server root location = //pbblsa304/app/bmc/BladeLogic/8.1/NSH/storage/
BLAST usage = Level 1 Enabled
PropertySync = Disabled
Hostname = PBBLSA303
OS Name = SunOS
Version = 5.10
Architecture = sparc
Number of Processors Available to JVM = 4
JRE Version = 1.6.0_13 from Sun Microsystems Inc.
JRE Data Model = 32-bit
JRE VM Name = Java HotSpot(TM) Server VM
Maximum JVM Memory (B) = 1,900,871,680
Total JVM Memory (B) = 1,900,871,680
Free JVM Memory (B) = 1,635,541,976
Used JVM Memory (B) = 265,329,704
Maximum File Descriptors = 8,192
Used File Descriptors = 291
FIPS 140-2 Mode = enabled
BLManager Current Time = Fri Dec 02 16:53:20 IST 2011
BLManager Started = Thu Nov 24 15:40:46 GMT+05:30 2011
BLManager PID = 16051
pbblsa303 (from Launcher on PBBLSA303)
Name = pbblsa303
Display Name = PBBLSA303
Server Type = ALL
State = VALID
Status = Ready
Elapsed Time = 8 day(s) 1 hour(s) 13 minute(s) 28 second(s)
Start Date = Thu Nov 24 15:39:51 GMT+05:30 2011
Needs Restart = false
Service Type = Automatic
There is single job server.
How much memory in the box?
How many cpus ? what class of sparc box is this ?
Box config is: 8 GB RAM, 1 CPU(4 cores) & it is 4u class.
The first thing that I would recommend to change is the work items number.
A work item is almost always bound to one target host, or to a component on a target.
So if you running a jobs against 100 targets it will be a good practice to change the number of work items to >100.
Second I would recommend using the “Process Spawner” (especially for Solaris).
If you are using 2GB RAM server it will be reasonable to configure the Process Spawner to 512 MB
It willl eliminate a "Peaks" in the Memory usage while starting new activity.
1 of 1 people found this helpful
The max heap you can configure on a 32bit jvm (what you are using) is 2g. the max work item thread count you can have there is 50. you should not configure a WIT count > 50.
instead of a single instance of the appserver (which i'm assuming you're running) you should create 1 CONFIG/NSH_PROXY and 2 JOB instances on your server. this will separate the java heap out so that the UI sessions should not be as affected by the compliance runs.
you should also enable the process spawner - having it running is not sufficient - you must run 'blasadmin -a set proc spawn true' and restart the appserver and spawner.
Yes, I can try this opton of adding one more app server instance.
As almost all configurations and integration with BBDSSA is done, Just want to know,
Do I need to do any other configuration modifications once I add appserver instance?
1 of 1 people found this helpful
You need to add a new config/nshproxy, two new job and remove the existing ‘default’ instance.
Does anyone know if yu can you remove the existing 'default' in vers 8.1.03?
Connect to another CONFIG instance in the env, run infra mgr, stop and delete the instance you want to remove.
I dont think so.
Why do you want to do that ?
Yes – you can. it’s just an appserver deployment, there is nothing special about it.
Yes, I realized so.