Skip to main content

I notice that every time I run Sync Check - with and without RUNCHKSUM(*YES) - all Sync Check jobs always run on Primary Node.   This puts workload burden - especially when RUNCHKSUM = *YES - on Primary node.    But when I think about this, it makes more sense to me that all Sync Check jobs should run on Backup node because, in practice, it is where the workload mostly or always is much lower than the Primary node.   I would like to ask if these is any control parameter that enables me to choose WHERE Sync Check jobs run?  I look into DMSETSVAL and find none.

Thanks.



------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

I notice that every time I run Sync Check - with and without RUNCHKSUM(*YES) - all Sync Check jobs always run on Primary Node.   This puts workload burden - especially when RUNCHKSUM = *YES - on Primary node.    But when I think about this, it makes more sense to me that all Sync Check jobs should run on Backup node because, in practice, it is where the workload mostly or always is much lower than the Primary node.   I would like to ask if these is any control parameter that enables me to choose WHERE Sync Check jobs run?  I look into DMSETSVAL and find none.

Thanks.



------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

Today, I try running Sync Check with checksum again and now I notice that, in the Backup node, there are also five CKB.... jobs running HAWORKER program in addition to the five CKP... jobs in the Primary node.   So, my question is no longer valid any more.

Thanks. 



------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

Today, I try running Sync Check with checksum again and now I notice that, in the Backup node, there are also five CKB.... jobs running HAWORKER program in addition to the five CKP... jobs in the Primary node.   So, my question is no longer valid any more.

Thanks. 



------------------------------
Satid Singkorapoom
IBM i SME
Rocket Forum Shared Account
------------------------------

Hi Satid,

Just to make sure anyone reading your posts understands your conclusion, sync checks jobs are run in pairs of source and target partner jobs.  These jobs compare the attributes of objects or in the case of *CHECKSUM sync checks, are comparing blocks of checksum results.  You can control how many parallel jobs are threaded to handle the workload and provide a balance of resource usage over time in the iCluster system values.

Regards,

Mark



------------------------------
Mark Watts
Sr. Software Engineer
Rocket Software Inc
Waltham MA US
------------------------------