ONTAP Discussions

Disk unpartition failed

boudjel38
27,202 Views

Hi everyone,

 

We are trying to remove all partitions on our netapp AFF300, to reinstall it with one root partition and 2 data partitions (as it was when we received it).

Right now we are in a configuration with one root partition and one data partition.

 

We followed this netapp documentation https://library.netapp.com/ecmdocs/ECMP1636022/html/GUID-CFB9643B-36DB-4F31-95D4-29EDE648807D.html

unsuccessfully.

 

We are strugulling with the command disk unpartition 0d.01.21. Here is the error :

Disk 0d.01.21 must not have file system partitions. All partitions must be spare
Do not proceed with unpartition if disk has file system partitions. Abort unpartition (y/n)? n
Jan 11 13:28:50 [localhost:raid.unpartition.disk.fail:notice]: Disk unpartition failed on Disk 0d.01.21 Shelf 1 Bay 21 [NETAPP X371_S163A960ATE NA50] S/N [S396NA0HA06398], error CR_DISK_WRONG_STATE, additional error info ().
disk unpartition: 0d.01.21 has partitions that are not spare.

 

We can't find any solution on google, so any help would be much appreciated 🙂

Regards

Jean-Pascal

1 ACCEPTED SOLUTION

marc_riedel
27,013 Views

Hi everyone,

 

this week I had a similar problem when I tried to re-install a cdot cluster. In my case it was a 2554 single-node system with SATA disks, which was shipped with ONTAP 9.0.

I booted the system and did "Install new software first" to load ONTAP 9.1. After this I went to maintenance mode to unpartition the disks und remove the ownership. I ran into the following error:

 

*> disk unpartition 0a.00.12
Disk 0a.00.12 must not have file system partitions. All partitions must be spare
Do not proceed with unpartition if disk has file system partitions. Abort unpartition (y/n)? n
disk unpartition: 0a.00.12 has partitions that are not spare.
Jan 19 10:18:43 [pfhn-cl3-01:raid.unpartition.disk.fail:notice]: Disk unpartition failed on Disk 0a.00.12 Shelf 0 Bay 12 [NETAPP   X306_HMRKP02TSSM NA00] S/N [P5JTUG5V], error CR_DISK_WRONG_STATE, additional error info ().

 

I got 4 disks which could not be unpartitioned and the reinstall did not went through.

 

Finally I netbooted ONTAP 9.0, went to maintenance mode and were able to unpartition the disks. After this the reinstall was possible.

 

it seems that in my case ONTAP 9.1 can not deal correctly with partitions created prior ONTAP 9.1 - So it would be interresting if a netboot of ONTAP 9 in your case would also solve the problem.

Then there would be an general issue in ONTAP 9.1 in combination with disk unpartioning.

 

Kind regards

 

Marc

View solution in original post

13 REPLIES 13

AlexDawson
26,696 Views

Hi there - there are no documented procedures for this, however I'd start with this:

 

Was HA disabled, both nodes halted, then one brought up in maint mode?

 

Then I would run "storage disk removeowner disk_name -data true" for all of the drives, and then "storage disk removeowner disk_name -root true" for all of them, then reboot and zero one node, then the other.

 

Let me know how it goes!

aborzenkov
26,679 Views

Clustershell commands are not available in maintenance boot, unless this has changed in 9.x.

AlexDawson
26,678 Views

That's a very good point!

 

Ok, so you'd unassign the data partitions in fully booted ONTAP (the presumption here being that you have no data on the data aggregates that you care about).

 

If that still doesn't let you unpartition the disk, I'd suggest the original poster open a support ticket - we might need to escalate to engineering to find the right mix of commands. The good thing about AFF is that reformatting is pretty quick 🙂

aborzenkov
26,676 Views

Do you see any leftover aggregates in "aggr status"? I wonder if this disk belongs to pool, not sure how pool is represented in maintenance mode.

 

As this disk is in external shelf, you probably can simply unplug shelf during installation and add it later.

marc_riedel
27,014 Views

Hi everyone,

 

this week I had a similar problem when I tried to re-install a cdot cluster. In my case it was a 2554 single-node system with SATA disks, which was shipped with ONTAP 9.0.

I booted the system and did "Install new software first" to load ONTAP 9.1. After this I went to maintenance mode to unpartition the disks und remove the ownership. I ran into the following error:

 

*> disk unpartition 0a.00.12
Disk 0a.00.12 must not have file system partitions. All partitions must be spare
Do not proceed with unpartition if disk has file system partitions. Abort unpartition (y/n)? n
disk unpartition: 0a.00.12 has partitions that are not spare.
Jan 19 10:18:43 [pfhn-cl3-01:raid.unpartition.disk.fail:notice]: Disk unpartition failed on Disk 0a.00.12 Shelf 0 Bay 12 [NETAPP   X306_HMRKP02TSSM NA00] S/N [P5JTUG5V], error CR_DISK_WRONG_STATE, additional error info ().

 

I got 4 disks which could not be unpartitioned and the reinstall did not went through.

 

Finally I netbooted ONTAP 9.0, went to maintenance mode and were able to unpartition the disks. After this the reinstall was possible.

 

it seems that in my case ONTAP 9.1 can not deal correctly with partitions created prior ONTAP 9.1 - So it would be interresting if a netboot of ONTAP 9 in your case would also solve the problem.

Then there would be an general issue in ONTAP 9.1 in combination with disk unpartioning.

 

Kind regards

 

Marc

mirko
25,888 Views

You saved my day :

indeed you must netboot with 9.0 and there the disk unpartition worked fine.

 

Extra hinte : My IIS first refused to serve the /netboot/kernel (file not found).  Added a mime-type in my IIS "*" as application/octet-stream

 

 

scottgelb
26,037 Views

There was a bug in 9.1RC2 for disk unpartition... 9.1GA fixes it though or revert to 9.0 for disk unpartition

 

High level procedure..zeroes out the system.

 

on each node..

--------------------

boot_ontap maint   # maint mode boot

aggr offline aggr0

aggr destroy aggr0

disk unpartition all

disk remove_ownership

disk show    # if any disks assigned, run "disk remove_ownership" again

halt

 

loader

--------

setenv allow-root-data1-data2-partitions? True  

saveenv

 

 

on node1 only

--------

boot_ontap menu   # menu boot

44/7     # hidden zero option

 

after completion... run option 4 on node1 then node2..

 

you will have the correct partitions...but I found on a larger systtem they are split across 3 shelves instead of 2...but 48 drives partitioned as expected.  The physical unparittioned drive assignment was not symmetric but that is easy to fix after...

 

Also note that if you have over 2 shelves in an SSD stack (2 is best practice, 4 is max) a best practice is to split disk assignment between nodes in the ha-pair instead of full disk shelf assignment per shelf

JKnorr
17,567 Views

I had the same issue in 9.1 GA. Only RC1 worked FYI.

KirillAntipov
25,976 Views

Hi!

Just today spent unforgettable 5 hours with partitions. AFF8080 was upgraded from 9.0 up to 9.1 and then partitions became un-unpartitionable. Error just the same:...disk has partitions that are not spare.

Solution in our case:

1. Reboot to maintenance both nodes. *>halt -> Ctrl+C -> maintenance (option 5).

2. Delete all aggregates (root too). *>aggr status -> *>aggr offline -> *>aggr destroy -> *>aggr status

3. Remove ownership.  *>disk remove_ownership all (several times. check *>disk show, it must be empty).

4. Assign all disks to one node. *>disk assign all

5. Reboot this node to maintenance from previous image (9.0). LOADER>boot_backup -> Ctrl+C -> maintenance (option 5).

6. Check ownership. *>disk show   *>disk show -n

7. Remove partitions. *>disk unpartition all

8. Reboot, boot ontap, create aggregates. LOADER>boot_ontap

 

Hope, this will help!

Regards

Kirill.

rshardha
25,833 Views

There is a bug in versin 9.1RC2, 9.1P1. If you want to remain on 9.1, then the you may go for 9.1RC1. This version allows you to delete the partitions.

JKnorr
17,566 Views

Agreed, I was only able to get 9.1 RC1 to work properly... GA and P1 gave me fits.

AlexDawson
17,517 Views

The BURT is detailed in http://mysupport.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=1057401, along with a bootloader recommendation to workaround

DER_OEST_
17,479 Views

Just stumbled upon this thread.

 

The bug is fixed in 9.1P2, which was released just days ago.

-> https://mysupport.netapp.com/NOW/download/software/ontap/9.1P2/

 

 

edit: I've just successfully removed ADP from some disks in order to move them to another cluster.

Workflow was something like this:

  • "system node migrate-root" to spare disks (they need to be zeroed or will be zeroed - takes some time)
  • "system node run -node xxx -command disk unpartition xx.xx.xx"
  • "disk zerospares"
  • "disk removeowner xx.xx.xx"
Public