Understanding Space

MammaGutt
Posts: 1577
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Understanding Space

Post by MammaGutt »

jwood.mls wrote:Alright, gotcha...so if you are running unmap and it does free up a lot of space, you probably should run compactcpg, but not something you would do all the time...

I did run unmap yesterday and it changed the used size on my VV's from 6% up to 23% (!!) in one case, so I'm guessing I do need to run the compact CPG to just get a clearer picture of where I am now.


That sounds weird. If you look at your service processor you should have historical data under files -> serial number -> timestampCONFIG if I recall correctly. Could you view that file from both before and after and paste the entire line for this volume from both files?
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
jwood.mls
Posts: 15
Joined: Thu Apr 17, 2014 11:58 am

Re: Understanding Space

Post by jwood.mls »

MammaGutt wrote:
jwood.mls wrote:Alright, gotcha...so if you are running unmap and it does free up a lot of space, you probably should run compactcpg, but not something you would do all the time...

I did run unmap yesterday and it changed the used size on my VV's from 6% up to 23% (!!) in one case, so I'm guessing I do need to run the compact CPG to just get a clearer picture of where I am now.


That sounds weird. If you look at your service processor you should have historical data under files -> serial number -> timestampCONFIG if I recall correctly. Could you view that file from both before and after and paste the entire line for this volume from both files?


I realize I worded that completely wrong. I ran unmap on 3 VV's. it REDUCED all of them. One of them it reduced 6%, one it reduced 10%, and one it reduced 23%. They all went down, so it did what it should have, right?
MammaGutt
Posts: 1577
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Understanding Space

Post by MammaGutt »

That sounds more like what I was expecting yes.
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
jwood.mls
Posts: 15
Joined: Thu Apr 17, 2014 11:58 am

Re: Understanding Space

Post by jwood.mls »

So after my unmap and then my compactcpg, here is what I have - definitely a difference:

Image

Image

From what I saw on another post here, it looks like that 1 GiB of failed is nothing to be concerned with?

Again, I really really appreciate everyone's assistance!
JohnMH
Posts: 505
Joined: Wed Nov 19, 2014 5:14 am

Re: Understanding Space

Post by JohnMH »

1GB failed is a media error on a single chunklet, from memory if 6 or more fail on the same disk the array will move the data and spare that disk out.
sanjac
Posts: 96
Joined: Thu Oct 26, 2017 1:21 am

Re: Understanding Space

Post by sanjac »

can you help me a bit...still confuzed about free space on my 3par

I've got 8200 with one VV one 3 CPG in AO.

System say:
Allocated 14.293 GiB
Free 12.491 GiB
Total 26.784 GiB

CPG1 (RAID1):
Used
3.400 GiB
Free
46 GiB
Allocated
3.445 GiB
Estimated usable free space
6.245 GiB

CPG2(RAID5):
Used
2.015 GiB
Free
15 GiB
Allocated
2.030 GiB
Estimated usable free space
10.375 GiB

CPG3(RAID6):
Used
2.304 GiB
Free
31 GiB
Allocated
2.335 GiB
Estimated usable free space
8.298 GiB

How much free space do I have? :9
JohnMH
Posts: 505
Joined: Wed Nov 19, 2014 5:14 am

Re: Understanding Space

Post by JohnMH »

It depends :-) and without knowing a bit about the makeup of the array or the disks each CPG points to I can only point you in the right direction.

Total 26,784 GiB - Allocated 14,293 GiB = Free 12,491 GB which is RAW capacity

However this free number is spread over 3 CPG's and it will depend on the parity overheads and whether the CPG's address the same or different disks as to how much you have available in each.

Note CPG space is expressed as usable (logical) based on the properties of the CPG not the RAW number.

CPG1 (RAID1):
Used 3,400 GiB (how much space you have written)
Allocated 3,445 GiB (how much space has been reserved by the CPG)
Free 46 GiB (how much space is available before the CPG grows again to accomodate new writes)
Estimated usable free space 6,245 GiB (this is usable space available for the CPG to grow into e.g. 6,245GB Usable * 2 for mirroring = 12,491GB Raw above)

CPG2(RAID5):
Used 2,015 GiB (Written)
Allocated 2,030 GiB (Reserved)
Free 15 GiB (Free until next grow)
Estimated usable free space 10,375 GiB (Estimated usable 10,375GB *1.2 for Raid 5(4+1) = 12,451GB Raw)

CPG3(RAID6):
Used 2,304 GiB (Written)
Allocated 2,335 GiB (Reserved)
Free 31 GiB (Free until next grow)
Estimated usable free space 8,298 GiB (Estimated usable 8,298GB usable *1.33 for Raid 6 (4+2) = 11,040GB Raw) - the remainder can possibly be explained by me getting my sums wrong or not being able to complete a full row based on the number of disks and set size being used.

Since the raid 1 raw figure matches your available free space I'd assume these CPG's are all sharing the same disks ? In which case all of the above are correct but it just depends on which one grows at any one time. But in summary you have 12,491GB RAW free of additional unallocated capacity that the system can allocate to any one of the CPG's based on a first come first served basis.

Try running a cli% showpd -c at the CLI and it'll show you the used and free chunklets available per disk, using the above you should be able to calculate the same.
sanjac
Posts: 96
Joined: Thu Oct 26, 2017 1:21 am

Re: Understanding Space

Post by sanjac »

JohnMN thank you for your answer....I was hoping you will writte somtihnig like YOU HAVE XXX GB FREE SPACE :D :D :D

Yes, all CPGs are using the same disks. RAW space doesn't meen a lot to me. Estimated usable free space seems more acurate data. But if CPG1 has 6,245 GiB of free space, I assume this number is even smaller bacause all CPGs are competing for the same space? :?
JohnMH
Posts: 505
Joined: Wed Nov 19, 2014 5:14 am

Re: Understanding Space

Post by JohnMH »

Yes potentially since they all point to the same disks they are all competing for the same space and depending where the growth occurs then each will incur different overheads e.g. 2x for raid 1, 1.2x for raid 5 etc.

TBH using AO on a single tier of disk is just making your job more complicated than it needs to be and probably for very little benefit.

The important thing is just to monitor the raw space and ensure it doesn't fall below a predefined level.
Post Reply