Extending a Pool?

[Art]

Hi,

I have an 8-bay FreeNAS 11.2 setup in RAIDZ1 and I just purchased 4 new harddrives to make the pool size larger.
These 4 drives are replacing existing drives so am replacing 1 drive at a time.

The issue is the supplier only shipped 3 of the 4 hard drives and may take a few weeks to ship the last hard drive.

So my question is, can I extend the pool after I have the 3 drives in and later extend again once I get the 4th hard drive replaces? Or should I wait till I have all 4 hard drives in before I extend the pool?

Thanks,

Art

 

[Kevin Horton]

You can replace three drives now, but the pool capacity can not increase until all drives have been replaced with higher capacity drives.

 

[Art]

Thanks for your reply, but to confirm your statement. Are you saying all 8 drives need to be replaced with higher capacity drives before I can increase the pool capacity?

 

[garm]

Yes

 

[sretalla]

If you haven't messed with your pool's autoexpand setting, it will just happen once your resilver is done on the last disk.

 

[Art]

Okay. Well since I currently do not want to buy 4 more disks of the same size my only other option I can see is to get a larger case and use the motherboards on board SATA connectors to add these 4 larger disks and create a new pool out of them.

So I currently have a Silverstone DS380 with 8x3.5" hard drives and 2x2.5" SSD.

Can you recommend a case that can support 12x3.5" and 2x2.5" slots that is still a smaller form factor?

 

[SweetAndLow]

What structure is your pool? How many vdevs? zpool status?

 

[Art]

My current pool status is Healthy and online.

Code:

[root@freenas ~]# zdb -U /data/zfs/zpool.cache
SSD:
    version: 5000
    name: 'SSD'
    state: 0
    txg: 16125037
    pool_guid: 3851491565436352485
    hostid: 2745325768
    hostname: ''
    com.delphix:has_per_vdev_zaps
    vdev_children: 1
    vdev_tree:
        type: 'root'
        id: 0
        guid: 3851491565436352485
        children[0]:
            type: 'mirror'
            id: 0
            guid: 7053917530877770641
            metaslab_array: 35
            metaslab_shift: 33
            ashift: 12
            asize: 998052462592
            is_log: 0
            create_txg: 4
            com.delphix:vdev_zap_top: 123
            children[0]:
                type: 'disk'
                id: 0
                guid: 16990541139416103981
                path: '/dev/gptid/01debfd8-d28c-11e5-9174-0cc47a7b6672'
                whole_disk: 1
                DTL: 121
                create_txg: 4
                com.delphix:vdev_zap_leaf: 129
            children[1]:
                type: 'disk'
                id: 1
                guid: 17010961523235980919
                path: '/dev/gptid/02136688-d28c-11e5-9174-0cc47a7b6672'
                whole_disk: 1
                DTL: 119
                create_txg: 4
                com.delphix:vdev_zap_leaf: 130
    features_for_read:
        com.delphix:hole_birth
        com.delphix:embedded_data
data:
    version: 5000
    name: 'data'
    state: 0
    txg: 17014085
    pool_guid: 9546008639022548689
    hostid: 2745325768
    hostname: 'freenas.local'
    com.delphix:has_per_vdev_zaps
    vdev_children: 1
    vdev_tree:
        type: 'root'
        id: 0
        guid: 9546008639022548689
        children[0]:
            type: 'raidz'
            id: 0
            guid: 13619103183019576364
            nparity: 2
            metaslab_array: 35
            metaslab_shift: 38
            ashift: 12
            asize: 47992182669312
            is_log: 0
            create_txg: 4
            com.delphix:vdev_zap_top: 306
            children[0]:
                type: 'disk'
                id: 0
                guid: 11878695458745325160
                path: '/dev/gptid/14b40bce-d22a-11e5-81f2-0cc47a7b6672'
                whole_disk: 1
                DTL: 296
                create_txg: 4
                com.delphix:vdev_zap_leaf: 307
            children[1]:
                type: 'disk'
                id: 1
                guid: 14341727756096334612
                path: '/dev/gptid/7abe931c-07da-11e6-9d16-0cc47a7b6672'
                whole_disk: 1
                DTL: 295
                create_txg: 4
                com.delphix:vdev_zap_leaf: 309
            children[2]:
                type: 'disk'
                id: 2
                guid: 1582421386450079856
                path: '/dev/gptid/160ff6d8-d22a-11e5-81f2-0cc47a7b6672'
                whole_disk: 1
                DTL: 294
                create_txg: 4
                com.delphix:vdev_zap_leaf: 310
            children[3]:
                type: 'disk'
                id: 3
                guid: 16835169984228047019
                path: '/dev/gptid/067bc6af-1cab-11e9-83c1-0cc47a7b6a5a'
                whole_disk: 1
                DTL: 440
                create_txg: 4
                com.delphix:vdev_zap_leaf: 128
            children[4]:
                type: 'disk'
                id: 4
                guid: 14146498870743979009
                path: '/dev/gptid/b4c94855-2146-11e9-83c1-0cc47a7b6a5a'
                whole_disk: 1
                DTL: 460
                create_txg: 4
                com.delphix:vdev_zap_leaf: 459
            children[5]:
                type: 'disk'
                id: 5
                guid: 14648671345942033649
                path: '/dev/gptid/0c660df6-20bc-11e9-83c1-0cc47a7b6a5a'
                whole_disk: 1
                DTL: 451
                create_txg: 4
                com.delphix:vdev_zap_leaf: 450
            children[6]:
                type: 'disk'
                id: 6
                guid: 12755236220142589614
                path: '/dev/gptid/18d7c1c9-d22a-11e5-81f2-0cc47a7b6672'
                whole_disk: 1
                DTL: 290
                create_txg: 4
                com.delphix:vdev_zap_leaf: 315
            children[7]:
                type: 'disk'
                id: 7
                guid: 16897148772954084853
                path: '/dev/gptid/3bb6c5f5-220f-11e9-83c1-0cc47a7b6a5a'
                whole_disk: 1
                DTL: 469
                create_txg: 4
                com.delphix:vdev_zap_leaf: 468
    features_for_read:
        com.delphix:hole_birth
        com.delphix:embedded_data

I already replaced 3 of the old harddrives and replaced them with the newer high capacity.

I do not know what you mean by "What Katie is your pool" - can you tell me how to find this?

 

[SweetAndLow]

What is your layout? How many vdevs? Is everything one big vdev? or do you have multiple vdevs?

 

[Art]

I have 2 pools each are a single vdev with multiple datasets defined.

Current Setup:
Pool: data
RAIDZ2
HD:
- 8x6 TB
vdevs: 1

Pool: SSD
MIRROR
HD:
- 2x512 GB
vdevs: 1

New Config:
Pool: data
RAIDZ2
HD:
- 4x6 TB
- 4x14 TB
vdevs: 1

Pool: SSD
MIRROR
HD:
- 2x512 GB
vdevs: 1

 

[Art]

Well I decided to go a different route all together after my misunderstanding on having to replace all current 8 hard drives to increase the storage capacity in the current vdev.

I am going to add a 12 bay SAS expander and a new HBA controller to support this extension. I currently have 9300-8i and I will replace with 9305-16i and add a back panel adapter to my case for the added 2xSFF-8644 external connectors.

Then I will purchase 2 more hard drives so I will have 6 new total and put them in the SAS expander.

So these new 6 hard drives will be a new vdev in RAIDZ2.

I assume I can add this new vdev to my existing pool and it will utilize the maximum space available on these larger hard disks; correct?

Also after this new vdev is added to the pool will the existing datasets automatically see the added storage to the pool? or do I need to do something to the datasets?

Are my assumptions correct or am I way off on my thinking? I (obviously) don't have much experience with managing pools/vdevs..

Thanks in advance :)

 

[Kevin Horton]

Well I decided to go a different route all together after my misunderstanding on having to replace all current 8 hard drives to increase the storage capacity in the current vdev.

I am going to add a 12 bay SAS expander and a new HBA controller to support this extension. I currently have 9300-8i and I will replace with 9305-16i and add a back panel adapter to my case for the added 2xSFF-8644 external connectors.

Then I will purchase 2 more hard drives so I will have 6 new total and put them in the SAS expander.

So these new 6 hard drives will be a new vdev in RAIDZ2.

I assume I can add this new vdev to my existing pool and it will utilize the maximum space available on these larger hard disks; correct?

Also after this new vdev is added to the pool will the existing datasets automatically see the added storage to the pool? or do I need to do something to the datasets?

Are my assumptions correct or am I way off on my thinking? I (obviously) don't have much experience with managing pools/vdevs..

Thanks in advance :)

As I understand it, three of your new large disks have already been used as replacements in the original 8 disk RAIDZ1 vdev. Your plan seems to imply that you will replace those disks with the original smaller disks. I have no idea if that is possible or not.

Assuming that it is possible to get those three disks out of the original vdev, your plan to expand the storage by adding a new 6 disk RAIDZ2 would work.

 

[Art]

I have put back the original (smaller) harddrives in the existing 8 disk RAIDZ2 vdev (originally thought it was RAIDZ1 - sorry). Now I am waiting to get a few more additional cables and the remaining larger hard drives and then I should finally be ready to go :)
Will post once all finished.

 

[Art]

Just an update. I added the new vdev 6x14tb RAIDZ2 in my SAS Expander and all is good. :)