Recalling DSN's from ML1 volume which does not show up in 3.

rahulbank · Posted: Tue Nov 04, 2008 4:40 pm

Hi All,

When I do an hsend list mcds on a dsn it shows migvol=online, this migvol is an ML1 volume, I am not able to see this dsn in3.4.

When I do an hsend fixcds d NBKMT1S.NOALT.D201008 display it shows the ml1 volume,how can i progress further on restoring it

This is some more detail from the command.
+0080 00000000 00000000 00000000 00000000 00000000 00000000 0000000F 40E2D4D4
* SMM*
+00A0 C9C74BC8 D4C9C74B E3F1F9F2 F6F0F94B C6C3C8D7 4BC6C3C8 4BD1F8F3 F0F64040
*IG.HMIG.T192609.FCHP.FCH.J8306 *

expat · Posted: Tue Nov 04, 2008 4:49 pm

If it does not show under 3.4 then the dataset is not catalogued.

Try to use HRECOVER from the backup sets.

rahulbank · Posted: Tue Nov 04, 2008 5:03 pm

HRECOVER I do not have a bcds for this dsn....I was trying to understand can we recover from an ml1 as it is showing online one more info when i hit hsend fixcds d NBKMT1S.NOALT.D201008 assignedbit(on) it shows an ml1 volume now..but still not in 3.4...

dick scherrer · Posted: Wed Nov 05, 2008 12:13 am

Hello,

rahulbank · Posted: Wed Nov 05, 2008 8:56 am

I am the storage guy

dick scherrer · Posted: Wed Nov 05, 2008 9:14 am

expat · Posted: Wed Nov 05, 2008 12:25 pm

As suggested previously, try to use the backup version to perform HRECOVER of the dataset.

Also in the past I have cataloged a dataset to volser MIGRAT which then allows HSM to recall. However this does not always work. You will need to relate the DSN against the HSM internal DSN and then list the VTOC of the ML1 volume (assumed to be DASD) and if it is there -OK. If not - the options are somewhat limited to none if HRECOVER has failed.

I suggest that you read, reread and execute the HSM AUDIT functions using NOFIX mode - unless you can guarantee absolutely no HSM activity during the time taken to perform the audit.

At least with NOFIX it will not attempt corrections for ongoing processes that have changed the CDS inflight.

rahulbank · Posted: Thu Dec 04, 2008 5:57 pm

I tried Define Nonvsam name devicetype volumes(migrat) and the dsn got cataloged in 3.4 but it will not recall as the hsm believes migratvolume = online and when recalling it says dataset not migrated...any suggestions...

enrico-sorichetti · Posted: Thu Dec 04, 2008 6:03 pm

if the dataset is mission critical You might

find out what a catalog entry for a migrated dataset looks like ( hex dump of the catalog records )

do the same for a generic dataset
run a define nonvsam for he dataset You are trying to recall

and using any zapping tool ( ditto should be able to do it ) fix the record

expat · Posted: Thu Dec 04, 2008 7:48 pm

You need to look at the MCDS record for the dataset and determine the HSM internal dataset name from the M (or is it A ?) record.

Then do a 3.4 on the ML1 volume that it thinks the internal dataset resides on. If the dataset is there - all well and good. Recatalog the dataset to volser=MIGRAT and then the RECALL should work OK.

If the internal dataset is not on the ML1 volume shown in the MCDS record, then the record is invalid and the dataset will never ever be able to be recalled. Which unfortunately from you last post appears to be the case

Hence my recommendation to perform the AUDITs on ALL of the CDS files. Find out what is wrong, and fix it, before it comes back and bites you on the butt !!!