Re: Mirroring by RRN slow for large file

Hi Pete,


Generally speaking the only time you should see latency on the apply side
is:

Resources. CPU/Memory
iTera was down for a time
Commitment Control
Locking of files - i.e other jobs running on the target for perhaps read
only functions
Large batch job generating many transactions in a short time.


The latency of data on the apply side shouldn't be happening but it
nothing to do with RRN processing vs keyed. I assume all of the access
paths are are *IMMED for update and none of them ate *DELAY.

EDTRBDAP will tell you if the system is rebuilding access paths. If it is
this would cause a transaction to queue but only until the access path was
re-built. Once it was then it should fly threw the data. I've seen upwards
of billions of transactions a day and no latency.

Has this always happened? Was there an event that may have caused this?
Have you ever done a role swap? or even a virtual role swap?

Product like iTera have been out for over 25 years now and have proven
methods of applying data. I suspect there is something else happening that
is causing the delay..

Thanks
Augie

Augie Palumbo
312-735-3723
702-826-2836

On Sun, Aug 13, 2017 at 8:02 AM, Nathan Andelin <nandelin@xxxxxxxxx> wrote:

Peter,

I haven't actually tested the performance of RRN vs. keyed index access on
IBM i, but the IBM i index search may actually return a RRN if it works
like some of the other DBMS products I've worked with. So a known RRN may
be the quickest way to find records.

I'm not that familiar with iTera, but I wonder if it may have a
configuration setting that controls the queue on the target system?

Nathan.


On Sat, Aug 12, 2017 at 11:39 PM, Peter Connell <Peter.Connell@xxxxxxxxxxx

wrote:

I assume that mirroring software (ITERA) that uses journals for
replication to another IBM i5 performs updates and deletes on the target
system by using the RRN stored in each journal to replicate the update or
delete.
This is a reasonable assumption since the mirroring software knows

nothing

about any index it might use to achieve the update of delete via an

access

path.

A large file on the source system of 170 million records which may

attract

several million updates and deletes in a day results in these being

queued

for long periods on the target system before the actual replication event
occurs to keep the file in synch.
I presume that the substantial latency observed between the source and
target updates and deletes is because the record be updated or deleted is
being located via RRN.
Perhaps this is because DB2 has to do a table scan to find the record
which is slower than an update via a unique index.

It's not clear if creating an index whose primary key is RRN would help
otherwise the only solution may be to do a custom replication that reads
the journal, and extracts the unique index to do an update or delete.

Regards, Peter

############################################################## This
correspondence is for the named person's use only. It may contain
confidential or legally privileged information, or both. No

confidentiality

or privilege is waived or lost by any mistransmission. If you receive

this

correspondence in error, please immediately delete it from your system

and

notify the sender. You must not disclose, copy or rely on any part of

this

correspondence if you are not the intended recipient. Any views expressed
in this message are those of the individual sender, except where the

sender

expressly, and with authority, states them to be the views of Equifax. If
you need assistance, please contact Equifax :- Australia
www.equifax.com.au/contact New Zealand www.equifax.co.nz/contact
##############################################################
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD