Way too ugly.
Paul
-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rob Berendt
Sent: Tuesday, September 20, 2016 4:02 PM
To: Midrange Systems Technical Discussion
Subject: RE: BRMS scatter saves
We could break it apart but not very easily.
WITH T1 AS (
select substr(x.objname, 1, 3) as grouper, sum(x.objsize) as receiversize from table (qsys2.object_statistics ('#MXJRN', '*JRNRCV')) AS X group by substr(x.objname, 1, 3)) SELECT T1.grouper, t1.receiversize from t1 order by t1.receiversize desc;
Datagroup DG receiver size
ALL 1,859,095,375,872
KRO 339,640,762,368
SYS 27,076,866,048
EDI 15,316,488,192
SMA 7,728,861,184
GDI 1,159,221,248
USR 147,914,752
MGR 147,259,392
391,217,373,184 = sum of all receivers besides the "ALL" ones 21.04% =80% of stuff is in "ALL"
We would have to break the "ALL" data group down. You have to be a little careful to ensure that one transaction stays within the same receiver. For example if you do a transaction which updates files in three libraries then all three of those libraries should be in the same receiver (in case you decide to start doing commitment control or some such thing). And if 90% of all the transactions within the "ALL" group are from ERPLXF anyway (that would involve further analysis) then you don't buy squat.
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From: "Steinmetz, Paul" <PSteinmetz@xxxxxxxxxx>
To: "'Midrange Systems Technical Discussion'"
<midrange-l@xxxxxxxxxxxx>
Date: 09/20/2016 03:10 PM
Subject: RE: BRMS scatter saves
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>
Rob,
Your largest library only contains journal receivers, correct?
Interesting?
No easy way to slit objects in this library when there all the same type,
and almost the same name?
ICOMS20350 *JRNRCV CABLEFILES ICOMS Journal Receiver - CB
ICOMS20351 *JRNRCV CABLEFILES ICOMS Journal Receiver - CB
Paul
-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rob
Berendt
Sent: Tuesday, September 20, 2016 3:01 PM
To: Midrange Systems Technical Discussion
Subject: RE: BRMS scatter saves
Thanks.
Our "big" library is journal receivers. We do INSANE retention of journal
receivers.
When you think of a restore strategy you should have your receiver library
restored before your data library. So you shouldn't do *ALLUSR (except
#BIGLIB) #BIGLIB
You could do
#BIGLIB
*ALLUSR (except #BIGLIB)
But do you want it there before QUSRSYS and QGPL?
But one or the other is the only work around if you want to scatter save
#BIGLIB.
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail
to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From: "Steinmetz, Paul" <PSteinmetz@xxxxxxxxxx>
To: "'Midrange Systems Technical Discussion'"
<midrange-l@xxxxxxxxxxxx>
Date: 09/20/2016 02:33 PM
Subject: RE: BRMS scatter saves
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>
Rob,
I remember discussing this in a previous thread.
I have a similar situation with one of my libraries, 5 times larger than
any other library.
The problem I have with IBM's solution is that you have to hardcode
libraries and objects within BRMS using multiple control groups.
I prefer not to do this, because every time a new library is created or a
library is deleted, you must revisit your BRMS control groups config.
I'm a strong advocate of *alluser, no changes needed to BRMS when new
library is created or a library is deleted, and you know you are backing
up 100%.
Prior to using *alluser, we were burned in the past when someone created a
new library, and it wasn't added to the save.
Gave you a vote.
Paul
-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rob
Berendt
Sent: Tuesday, September 20, 2016 1:34 PM
To: Midrange Systems Technical Discussion
Subject: Re: BRMS scatter saves
Sounds familiar, like I've discussed this before.
Anyway, I opened up an RFE and if this pertains to you then I'd appeciate
your vote.
https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe&CR_ID=94797
Rob Berendt
midrange.com
