Odd observation

Status
Not open for further replies.

KMR

Contributor
Joined
Dec 3, 2012
Messages
199
Hey folks,

I've been playing with a FreeNAS VM in ESXi as a sandbox to familiarize myself with FreeNAS operations before attempting them on my real build and I ran across something odd. I configured a FreeNAS VM with 8 virtual disks that are each 8GB (results in 8.6GB disks presented to VM) apart from the OS disk. I was playing around configuring them in RAIDZ2 and RAID 10 pools and found that if I created a 4 disk RAID 10 pool through the GUI I got a pool that was 11.8GB and if I crated that same pool on the command line the pool size was 15.6GB. Is there any reason for the variation? I assumed it had something to do with the virtual disks but considering they are all the same size I thought it was odd.

Thanks,
 

gpsguy

Active Member
Joined
Jan 22, 2012
Messages
4,472
The "odd observation" is exaggerated, due to the size of your virtual disks.

FreeNAS reserves 2Gb on each disk for swap. You can see it, with the "gpart show" command.
 

KMR

Contributor
Joined
Dec 3, 2012
Messages
199
Yes, but shouldn't the pool sizes be the same within the VM? I have 4 virtual disks in RAID 10 with a pool size of 11.8GB and 4 more virtual disks in RAID 10 with a pool size of 15.6GB with the only difference being the way they were created. Both pools exist at the same time and all disk sizes are the same. The 2GB swap size is the same across all of the 8 disks.
 

KMR

Contributor
Joined
Dec 3, 2012
Messages
199
Code:
[root@freenas] ~# zpool status
  pool: iscsi
 state: ONLINE
  scan: none requested
config:

        NAME                                            STATE     READ WRITE CKSUM
        iscsi                                           ONLINE       0     0     0
          mirror-0                                      ONLINE       0     0     0
            gptid/c250e4fd-d7b7-11e2-b62b-000c29f37b8c  ONLINE       0     0     0
            gptid/c278d3cb-d7b7-11e2-b62b-000c29f37b8c  ONLINE       0     0     0
          mirror-1                                      ONLINE       0     0     0
            gptid/26fa59f9-d7bb-11e2-b62b-000c29f37b8c  ONLINE       0     0     0
            gptid/271f0b44-d7bb-11e2-b62b-000c29f37b8c  ONLINE       0     0     0

errors: No known data errors

  pool: test
 state: ONLINE
  scan: none requested
config:

        NAME        STATE     READ WRITE CKSUM
        test        ONLINE       0     0     0
          mirror-0  ONLINE       0     0     0
            da1     ONLINE       0     0     0
            da2     ONLINE       0     0     0
          mirror-1  ONLINE       0     0     0
            da3     ONLINE       0     0     0
            da4     ONLINE       0     0     0

errors: No known data errors

 

paleoN

Wizard
Joined
Apr 22, 2012
Messages
1,402
The 2GB swap size is the same across all of the 8 disks.
No, it isn't. Your second, manually created pool is using the whole disks and not the second partition like the GUI pool. Even if da1-da4 are partitioned you aren't using the partitions.
 

KMR

Contributor
Joined
Dec 3, 2012
Messages
199
Awesome, thanks for clearing that up. I found those steps online and was wondering why I got that result. I am planning on adding a 4 disk RAID 10 pool to my "production" server (after backing everything up!) and wanted to do some testing before hand. I will be adding the new pool as an iSCSI device extent for an ESXi datastore and didn't want to muck anything up.

Out of curiosity (no, I'm not planning on doing this) what would be the potential problems in not using the swap on a system such as mine?

As always, thanks for the help!
 

paleoN

Wizard
Joined
Apr 22, 2012
Messages
1,402
Out of curiosity (no, I'm not planning on doing this) what would be the potential problems in not using the swap on a system such as mine?
The system should never be using swap. If it is you have sized something wrong.

IMO, the main caveat with using whole disks is then attempting disk replacement via the GUI. The GUI will partition the replacement drive. Which by default will make it too small. If you were to use whole devices then manage pool from the CLI.

I've also seen some FreeBSD heavyweights recommend to always configure swap as this can impact the FreeBSD VM subsystem. Of course I've seen a few say they don't bother with any swap at all. Personally, I configure some swap that the system never uses.
 
Status
Not open for further replies.
Top