I think I have this incorrect, the system I thought I saw it enabled I may have been looking at the overall file deduplication settings and a colleague has told me he has found a white paper stating that deduplication cannot be enabled on manually created filesystems (there is a block pool underneath). Is that correct?
That's what they're getting when they try to enable it.
Just to clarify, we created a file pool on an underlying block pool by manually creating the block pool and carving out the luns and putting them into the file storage group.
We did this as we had issues in the past with 'wonky' performance from files system that have been upgraded several times and the thought was using the horizontal redistribution of a block pool would keep the performance balanced.
During the deduplication process, the file system must have enough free space available that is equivalent to the size of the original file to be deduplicated, plus the size of the compressed version of the file to be stored:
An additional 1 percent of the file system must be free.
If auto-extension is enabled, additional 1 percent below the auto-extension threshold must be free.
The system in question is brand new with no file data on there at all, so where it states 'free space' are we talk about the capacity that's been allocated to the file system or the underlying file pool?
I hope that you have this as RAID Group dedicated to the FS. Moreover EMC doesn't recommend to enable compression on FS, so just keep in mind. This is the best practice guide :
it means free space *inside* the file system that you are deduplicating.
as with compression on other systems the VNX needs some space to store the compressed file before it removes the uncompressed version. plus some space to store the hashes for SIS.
Rainer
P.S.: please refer to EMC docs using emc.com links and not re-publish to other services unless you have the Copyright permission to do so
Thanks for the tip, I think we've found the issue, from what I've read you cannot enable auto-extension on a manually created file system and deduplication can only be enabled on a file system with it enabled.
edhoward
2 Intern
•
214 Posts
0
August 20th, 2014 06:00
Here you go:
Filesystem must be created from a storage pool to be able to extend.
Message ID: 13421841491
It is built on a storage pool.
Thanks,
Ed
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 20th, 2014 06:00
Well, you missed to mention what actual error message you are getting and what the data mover log says
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 20th, 2014 07:00
The message you quoted has nothing to do with VNX file deduplication
dynamox
9 Legend
•
20.4K Posts
0
August 20th, 2014 07:00
when you run "nas_fs -info " ..under pool value what does it say ?
edhoward
2 Intern
•
214 Posts
0
August 20th, 2014 07:00
I think I have this incorrect, the system I thought I saw it enabled I may have been looking at the overall file deduplication settings and a colleague has told me he has found a white paper stating that deduplication cannot be enabled on manually created filesystems (there is a block pool underneath). Is that correct?
edhoward
2 Intern
•
214 Posts
0
August 20th, 2014 07:00
That's what they're getting when they try to enable it.
Just to clarify, we created a file pool on an underlying block pool by manually creating the block pool and carving out the luns and putting them into the file storage group.
We did this as we had issues in the past with 'wonky' performance from files system that have been upgraded several times and the thought was using the horizontal redistribution of a block pool would keep the performance balanced.
dynamox
9 Legend
•
20.4K Posts
0
August 20th, 2014 07:00
i think there is some confusion, block pool is different from the pool that is created within datamovers on top of which you create file systems.
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 20th, 2014 08:00
I would suggest to look at the manual for VNX file deduplication – it does spell out the prerequisites
My guess – your file system is simply too full
edhoward
2 Intern
•
214 Posts
0
August 20th, 2014 08:00
In this document it states (page 9):
During the deduplication process, the file system must have enough free space available that is equivalent to the size of the original file to be deduplicated, plus the size of the compressed version of the file to be stored:
An additional 1 percent of the file system must be free.
If auto-extension is enabled, additional 1 percent below the auto-extension threshold must be free.
The system in question is brand new with no file data on there at all, so where it states 'free space' are we talk about the capacity that's been allocated to the file system or the underlying file pool?
Thanks for all your help so far,
Ed
Anonymous User
375 Posts
0
August 20th, 2014 21:00
I hope that you have this as RAID Group dedicated to the FS. Moreover EMC doesn't recommend to enable compression on FS, so just keep in mind. This is the best practice guide :
https://www.emc.com/collateral/white-papers/h12209-vnx-deduplication-compression-wp.pdf
Thanks
Rakesh
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 21st, 2014 02:00
it means free space *inside* the file system that you are deduplicating.
as with compression on other systems the VNX needs some space to store the compressed file before it removes the uncompressed version. plus some space to store the hashes for SIS.
Rainer
P.S.: please refer to EMC docs using emc.com links and not re-publish to other services unless you have the Copyright permission to do so
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 21st, 2014 02:00
You are confusing block compression with file compression
edhoward
2 Intern
•
214 Posts
0
August 21st, 2014 03:00
Thanks for the tip, I think we've found the issue, from what I've read you cannot enable auto-extension on a manually created file system and deduplication can only be enabled on a file system with it enabled.
Thanks for your help,
Ed
Rainer_EMC
4 Operator
•
8.6K Posts
0
August 21st, 2014 03:00
You dont need fs auto-extention for fs deduplication to work – you just have to have some free space in it
Anonymous User
375 Posts
0
August 21st, 2014 05:00
Hi Rainer,
What I meant is that there should be compression on that RAID Group, from where the file space will be allocated.
Thanks
Rakesh