|
View previous topic :: View next topic
|
| Author |
Message |
kalaikala
New User
Joined: 21 Jul 2007
Posts: 27
Location: Chennai
|
|
|
|
How to Recall the set of sequential files and members which is in migration state. Now i have a rexx script in which i have to do each and every file one by one(I cant do it continuously some time gap is needed between each submission). And also i need to maintain the information that which file is successful and which is not. This also i am doing manually  , so it consumes lot of time. If any one knows simple way to do this please help me out. |
|
| Back to top |
|
 |
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
| Why not put all the files into one rexx and run it once |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
If you do this
| Quote: |
| Why not put all the files into one rexx and run it once |
and execute the rexx in batch, you will have all of the info in a sysout dataset for use and re-use later. |
|
| Back to top |
|
|
PeD
Active User
Joined: 26 Nov 2005
Posts: 459
Location: Belgium
|
|
|
|
One advice also : writing a small REXX exec is easy but in that case ( specially with generic name ) most of the times you will recall more files that you need.
You will then keep resources ( tapes robots, .. ) and you work not alone.
I saw on monday morning or back from holidays some guys making an programmed exec and recalling more than 4.000 data sets.
Others have just to wait even for one recall process !!!! |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello Kalai,
Please clarify this:
| Quote: |
| set of sequential files and members |
Is there something special going on with "members" (the recall would bring back the entire pds)?
Just how many files are going to be recalled (probably my misunderstanding, but i had a fairly small number in mind - Thanks for the reminder, Pierre). If you need to do this for thousands of datasets, you may not be able to fit all of them onto the current dasd, not counting the impact on other processing.
I also notice that part of your task is to keep track of which ones could be recalled and which failed. Do you expect recall failures? Sounds very scary.
If you have many. many datasests to verify, i'd suggest breaking them into managable groups and run those in batch. Between recall runs, force migrate the ones that were successful and not immediately needed for processing - you could use the same list of dsns used to do the recall leaving out any failures. That way your process would be mostly automated and not cause resource problems. |
|
| Back to top |
|
|
kalaikala
New User
Joined: 21 Jul 2007
Posts: 27
Location: Chennai
|
|
|
|
Hello Friends,
Thanx.  I ll try your suggestions and let you know.
I have to this processes nearly 500 files on every Monday. So nearly half a day is wasting. And i am very new to mainframe so it ll take time for me to implement it. |
|
| Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
By submitting them all in one go is more efficient from a HSM point of view.
HSM will mount eash ML2 tape only once for all of the requests, so if you break them down into batches, you may end up mounting the same volumes numerous times. |
|
| Back to top |
|
|
PeD
Active User
Joined: 26 Nov 2005
Posts: 459
Location: Belgium
|
|
|
|
| Quote: |
| By submitting them all in one go is more efficient from a HSM point of view. |
Hi Martin,
except it can be rules putting in low priority the one asking for many recalls and letting go first the one(s) whith small amount of recalls.
It could be the case ( it is the case ) at the shop you are working  |
|
| Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
Hi Pierre,
Mmmmmmmmmmmmmmm,
That must be one hell of an exit / code change to get inside of HSM like that.
Or do you mean a monitor that dynamically adjusts the DP of the job executing to load the requests into the HSM queue.
P.S. - Must be my turn to buy lunch  |
|
| Back to top |
|
|
kgumraj
Active User
Joined: 01 May 2006
Posts: 151
Location: Hyderabad
|
|
|
|
Hi Kalaikala,
Everyone has same problem, whether you are new to mainframe or experiened. But we need to be little care when handleing situations.
What I suggest is
1. You can use IEFBR14 program to recall quickly.
2. Or use HRECALL in JCL.
3. With my little exp I have found ".CNTL" files dont migrate as other files do. Check this.
Thanks |
|
| Back to top |
|
|
murmohk1
Senior Member
.jpg)
Joined: 29 Jun 2006
Posts: 1436
Location: Bangalore,India
|
|
|
|
Kiran,
| Quote: |
| With my little exp I have found ".CNTL" files dont migrate as other files do. Check this. |
What do you mean by .CNTL files? |
|
| Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
| The management of datasets, based on HLQ, LLQ, or various other dataset naming combinations are all site specific, depending on how the SMS environment has been set up by the storage group. |
|
| Back to top |
|
|
kgumraj
Active User
Joined: 01 May 2006
Posts: 151
Location: Hyderabad
|
|
|
|
Expact,
I didnt know it was Shop specific. Thanks.
Murali,
.CNTL files are Control files. usually we specify all Source, JCL, CLIST datasets as
USERID.JCL.CNTL etc.. |
|
| Back to top |
|
|
ug123
Warnings : 1
New User
Joined: 01 Mar 2007
Posts: 39
Location: Chennai
|
|
|
|
Hi Kalai,
Could u please post the code which u are using now.. because i too need it..
Thanks,
Ugandar |
|
| Back to top |
|
|
|
|
