This post is more than 5 years old
1 Rookie
•
50 Posts
0
2842
June 10th, 2008 16:00
please explain the differience "symcfg refresh" vs "symmask refresh"
Question from a newbie:
when to use "symcfg discover",
does this affect SYMAPI database .bin file on local symcli host?
when to use "symmask refresh", what this affect?
Thanks in advance.
Jason
when to use "symcfg discover",
does this affect SYMAPI database .bin file on local symcli host?
when to use "symmask refresh", what this affect?
Thanks in advance.
Jason
No Events found!



xe2sdc
4 Operator
•
2.8K Posts
0
June 12th, 2008 01:00
The DMX stores its configuration (hypers, metavolumes, frontend mapping and a lot of many different things) in its binfile.
The VCMDB is a very specific database that contains informations regarding hosts and their relation with the frontend (masking, lun offset, DLA, hba_flags).
With the symconfigure command you are changing the binfile. With symmask/symmaskdb you query (and change) the VCMDB. They have little if no relation. You can mask whichever volume you want, even granting a given WWN access to a metamember.
When you isse "symmask refresh" you are simply asking the frontend adapters to reload their copy of the VCMDB from the actual database. That will enable any change made to the VCM after any previous refresh. The hosts will always "see" the updated VCM (if I run a query from a different host I can see changes made BEFORE committing them).
When you commit changes via symconfigure you are changing a different database (the binfile) and you are simply changing the way your storage is configured anywhere from the backend up to the mapping on frontend. Sometime you can automagically mask devices while you map them .. I don't use this feature very often .. mostly becouse when you automagically mask from symconfigure you can give only a single WWN while I usually have to mask devices to multiple wwns.
In your scenario, the second change won't be allowed since metadevices are configured in binfile. Both changes (first one, when you create meta, second one when you try to use again 00B5) will try to modify the binfile. And the second change will find that 00B5 is already part of a meta and will fail. No VCM here
dynamox
9 Legend
•
20.4K Posts
0
June 10th, 2008 17:00
jasonj1
1 Rookie
•
50 Posts
0
June 11th, 2008 14:00
adding/removing devices to/from WWNs ..and then after
of your remove command your would run symmask -sid
123 refresh to actually commit the changes to FAs.
thanks for the explanation. more question regarding "symmask refresh".
you say "symmask refresh" used to commit the change to VCMDB. so
assuming I have hostA with symCLI installed, and hostB with ECC installed.
both talk to the say array.
1. after I issue "symmask add dev .." on hostA, but without "refresh", when will the change be commited to VCMDB? is there a timer of some sort that auto commit for you?
2. Is the change only reflect on the port(s) that hostA connect to ? or will the commited VCMDB automatically reflected on ECC server, that may connect through differient FA port(s)? Does hostB also need some kind of refresh
to see the VCMDB changes?
3. is the way VCMDB operatechanged from MCode 5670/5671 or later ?
Thanks again.
Jason
dynamox
9 Legend
•
20.4K Posts
0
June 11th, 2008 19:00
2) when you run symmask refresh, changes to VCMDB are reflected globally, all FAs are immediately aware of the changes to VCMDB. As soon you as you commit changes on HostA, when HostB attempt to do any kind of VCMDB query ..it will see the changes that were commited by HostA. HostB is accessing the same VCM database only through a different FA.
3) not that i am aware of.
xe2sdc
4 Operator
•
2.8K Posts
0
June 12th, 2008 00:00
5670/5671 or later ?
Can you please explain why you feel that something changed in VCM between 5670 and 5671 ??? Things changes slightly between code releases/patches .. however if you describe what you've seen maybe we can explain what changed
VKR-f4y7L
17 Posts
0
June 12th, 2008 01:00
I was under the presumption that symmask refresh , refreshes the contents of VCMDB so that all hosts can see the updated d/b but the actual commit is done as soon as we run the symconfigure with a commit .
For eg i create a meta 00B6 using devices 00B5 and 00B6 through symmconfigure command and i dont refresh the VCMDB by running the symmask refresh .
After say 2 days i try to create another meta using the device 00B5 .
Considering the above scenario , i think symm would not allow me to perform the action since the meta 00B6 has already been created even though the VCMDB is not committed .
Please let me know if i am wrong in my thinking on this .
Regards ,
VKR
VKR-f4y7L
17 Posts
0
June 12th, 2008 02:00
Crystal Clear now
I appreciate your patience in explaining each and everyone n no of times the very basic concepts .
xe2sdc
4 Operator
•
2.8K Posts
0
June 12th, 2008 03:00
RRR
4 Operator
•
5.7K Posts
0
June 12th, 2008 03:00
RRR
4 Operator
•
5.7K Posts
0
June 12th, 2008 03:00
As a matter of fact: I was outside mowing the grass at the time 11 orange colored grown up men were running behind a leather ball.
xe2sdc
4 Operator
•
2.8K Posts
0
June 12th, 2008 03:00
Here you can throw any question (from very basic to advanced ones) .. and hopefully someone will answer
dynamox
9 Legend
•
20.4K Posts
0
June 12th, 2008 03:00
a query from a different host I can see changes made
BEFORE committing them).
Thank you for correcting me.
so when you backup your VCM database, are you backing up what's loaded in memory on FAs or just the copy of VCM database that could contain entries that have not been committed to FAs ?
MarcT2
2 Intern
•
131 Posts
0
June 12th, 2008 04:00
RRR
4 Operator
•
5.7K Posts
0
June 12th, 2008 04:00
xe2sdc
4 Operator
•
2.8K Posts
0
June 12th, 2008 04:00
up what's loaded in memory on FAs or just the copy of
VCM database that could contain entries that have not
been committed to FAs ?
The "original" copy is the one on the disk .. When you refresh the VCMDB you are actually refreshing its copies that each FA keeps in its memory.
All symm* commands do modify the "original" copy .. when you updated the "original" copy you ask the FA to reload their copies