AIX : JFS Filesystem and Big VG

Hi Experts,

I would like to create a Big Volume Group with the filesystems as attached excel sheet.

Could you all pls advice what's the appropriate PP size I should put when configure the Big VG?

I heard that there's mbpi limitation for JFS.

Thank you very much and have a nice day.

Best Regards,
Terrence Tan
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Hi again,
why would you use a BIG VG and jfs?
Use a SCALABLE VG and jfs2 and there's virtually no limit.
You will have no limit for PPs per PV, and "enough" PPs per VG (default is 131072) , so you can freely choose your PP size. I'd recommend 128-256 MB, maybe 256-512 MB for the two big ones,  .... vgds002 and .... vgds005.
Inodes are allocated dynamically with jfs2, so no problem with nbpi (that value doesn't exist in jfs2).
If you have to stay with Big VG and jf2, please let me know. I'll help you.
terrencetanAuthor Commented:
Dear woolmilkporc,

I'm sad to say that the customer choose to use Big VG and JFS filesystem.

I might need your advice on the PP size of VG for all the VGs with filesystems size as attached doc.

Thank you very much and have a nice day.

Best Regards,
Terrence Tan
sounds not very sensible, but anyway ...
Sorry for not having asked before, but I need to know the maximum sizes of the physical disks that make up your volume groups (now and with possible extensions in the future) for the calculation of supported PP sizes.

Learn SQL Server Core 2016

This course will introduce you to SQL Server Core 2016, as well as teach you about SSMS, data tools, installation, server configuration, using Management Studio, and writing and executing queries.

,,, and one more question: Will there be many small files or only rather few big ones un the new filesystems? Please tell me on a per-VG basis!
terrencetanAuthor Commented:
Dear woolmilkporc,

Actually for the Production Server filesystems are share filesystems and Shared VG (HACMP)
The VGs consists of multiple EMC LUN disks. Each LUNs are 17GB. (veri weird LUN Size - but Japanese standard). All this filesystems will be another bcv copy which will mounted on the Dev Server.

On the Dev server, all filesystems are VGs are not cluster filesystems.

Any big files in the filesystems so far i hav no idea. But I already informed customer max file size only 64GB and customer guaranteed there will not hav file size exceed this limit.

This is really first time to meet this kind of weird situation. Really need ur expertize to advice what's the right PP size of the VG.

Is it better combination for Scalable VG and JFS filesystem? Does this scalable VG supported in HACMP environment?

Thanks in advanced.

Best Regards.
Terrence Tan
Hi Terrence,

what I understood so far:

1) each LUN will only be 17 GB in size
2) VG sizes range from 130 GB to 1.370 GB
3) files are not huge, but also not tiny (not only 10K-100K or so)
4) some of the VGs will run with HACMP shared/enhanced-concurrent

1) means that we'll not hit the 1016 PP per PV limit with a PP size > 16M. But it means, too, that with the BigVG limit of 128 physical volumes you can never have a VG > 2.176 GB (unformatted).
2) means that with the biggest VG size of 1.370 GB we will need a PP size of at least 64 MB, making nearly 22.000 PPs for that VG. This gives good granularity, but also gives much overhead in partition table processing, so I'd tend to choose 256 MB as the PP size (at least for the big one).
3) means that we could live with the standard jfs nbpi setting of 4096, although 32768 or 65536 might be better, to save space (less inodes = more space for data).
You will most probably have files > 2GB, so you will need to make the filesystems "large file enabled" (bf=true).
4) HACMP shared/enhanced-concurrent is supported with any VG type (standard/big/scalable) and also with jfs as well as jfs2.

That said, my recommendations:

- VG: PP size 256 MB for "aybq53-vgds005", 128 MB for the rest (but you could take 256 MB for all VGs, to keep consistency)
- jfs: nbpi 32768, bf=true, agsize 64 (default)

Scalable VG in combination with jfs will only raise the 2.176 GB per VG limit. All other considerations above stay valid, yet scalable VG will allow for much more growth.

jfs2, on the other hand, will make the considerations regarding nbpi and bf=true unnecessary.

There's no sound reason for big VG and jfs, as far as I can tell. Scalable VG and jfs2 are far better.
Just keep in mind to follow IBM's recommendation of not using jfs2's inline-logging with HACMP.

Good luck!


Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Unix OS

From novice to tech pro — start learning today.