About aborzenkov

aborzenkov · ‎2015-10-08

How can I obtain license codes for ISCSI, nfs and so forth, for my second-hand NetApp FAS2040? You have to ask original owner to give them. And I do not imply that you can legally use them even in this case ... IANAL. When I install 3 disks and install Ontap 7.3.2 RC1, will it have licenses? No. You will need to manually register them. And you will need original installation package of Data ONTAP to finalize creation of root volume, because some content is simply extracted from this package and is not present on boot medium.

aborzenkov · ‎2015-10-08

The other KB you mentioned covers the opposite scenario of the one I'm doing (it covers converting cdot filers to 7-mode). Oh, this is just how search on KB works. The article you mention is https://kb.netapp.com/support/index?page=content&id=1014786 but it has text 1015457 in its body and so is found during search. I assume you do not have permissions to access 1015457 (or are doing search without logging in); otherwise it prefers direct KB number match. Full URL is https://kb.netapp.com/support/index?page=content&id=1015457 - can you acces it?

aborzenkov · ‎2015-10-07

Please, do not add every possible tag just because they are available. You cannot have C-Mode and 7-Mode at the same time. I assume you have 7-Mode. To expand aggregate you need free disks. Either you free some or buy some. There are no miracles. When movin root volume all configuration is kept as long as no changes happens between copying of root volume and booting from new root. Make sure to schedule it with minimum delay. Theoretically it is possible that filer decides to change domain password at exactly the wrong time but that is easily fixed by resetting password.

aborzenkov · ‎2015-10-07

During takeover a virtual instance of partner is started (compare with vfiler). So you still have your FilerA with all its configuration including local users.

aborzenkov · ‎2015-10-07

If not [creating files], then you should be fine without. Then why NetApp makes it mandatory to turn ucode on when LUN is created on volumes?

aborzenkov · ‎2015-10-06

Was file system size also increased? What df -h says?

aborzenkov · ‎2015-10-05

Root volume can be relocated non-disruptively using takeover/giveback; this is the last thing to worry about 🙂

aborzenkov · ‎2015-10-04

Assuming defaults on both NetApp and server sides it normally is expected to "just work" when you connect them. Do you observe any issues?

aborzenkov · ‎2015-10-02

It has been replaced with KB 1015457; can you access it? I am not sure what "hardware level HA is enabled" means, but I suspect it simply tries to say that only HA pairs are supported in multi-node clusters. I.e. you can only add whole HA pairs (both nodes), not individual nodes. Otherwise while repurposing controllers is really easy, there could be some unexpected issues depending on previous course of events. So it is always preferred if someone with good NetApp experience does it.

aborzenkov · ‎2015-10-01

Yes, it should be possible, but you are advised to check release notes. Anyway, update to 7.3.7 takes 15 minutes and does not harm. No new licenses are required for 8.1.

aborzenkov · ‎2015-10-01

Connect null-modem serial cable, configure your terminal as 9600, 8 bit, no parity, switch FAS2040 - it should attempt to boot. If it does attempt to boot, you can then initialize disks; depending on version currently installed you may need to have full operating system archive. It could be a challenge to work with it if you do not have access to NetApp support site downloads ...

aborzenkov · ‎2015-09-30

Single node cluster is still an option ...

aborzenkov · ‎2015-09-30

I think it depends on where those directories are located and whether their content is overlapping or not. You can specify multiple servers and search DNs. For search DNs it should continue if nothing is found in the first one, but as far as I understand multiple servers are for failover. I.e. if test and production are on different servers it probably cannot be done.

aborzenkov · ‎2015-09-30

What makes you believe you have any other relationship in the first place? list-destinations is expected to show all relationships with current cluster (SVM) as source. There is nothing wrong for them to have current cluster as destination.

aborzenkov · ‎2015-09-29

This depends on which Data ONTAP version and licenses you have and how much downtime you can afford. Sync mirror is zero downtime, but IIRC it is not supported between different vendors. Vol move is zero downtime but works for LUN only. Snap mirror offers the least downtime but requires license. Otherwise vol copy works always but means no client access for the copy duration. Other considerations are whether you have existing vault relationships etc. pulling in professional service may be the best approach.

aborzenkov · ‎2015-09-28

Well, while it is certainly unusual, I do not see anything wrong in having multiple interfaces with cluster management role. Of course, there could be just one routing table for admin SVM, so you really can only access one of them from different subnet.

aborzenkov · ‎2015-09-28

Please paste output of "network interface show".

aborzenkov · ‎2015-09-25

No. You cannot remove disks from aggregate. You have to destroy the whole aggregate. If you have data (volumes) on it, you need to relocate them to another aggregate. If the only volume left is the root one you can simply reinitialize this filer from special boot menu. It will create 3 disks RAID_DP aggregate and leave all other disks as spare.

aborzenkov · ‎2015-09-24

Does http://community.netapp.com/t5/Data-ONTAP-Discussions/SnapMirror-to-SnapVault-or-SnapVault-to-SnapMirror/m-p/107345#M22150 help?

aborzenkov · ‎2015-09-23

It is technically possible, but I do not think it is supported.

aborzenkov · ‎2015-09-23

You can add nodes, but you cannot merge two clusters and keep data and SVM of both. Nodes will need to be wiped clean, so any data will be lost. To preserve data it has to migrated to new cluster first.

aborzenkov · ‎2015-09-22

That's correct. With 8.2 you can only mirror data; but SVM must be created and configured on remote site. Also all configuration (shares, policies, user mapping etc) must be done manually on remote side. 8.3.1 introduced SVM replication so configuration is done automatically. You may look into it.

aborzenkov · ‎2015-09-22

Deleted volumes are added to volume recovery queue and removed after 12 hours. You can change it using vserver modify -vserver <vserver> -volume-delete-retention-hours or explicitly remove queued volumes using volume recovery-qeueue purge.

aborzenkov · ‎2015-09-21

vol options volume no_atime_update on

aborzenkov · ‎2015-09-21

Why do you need to rename them in the first place?

Re: How to bring a second-hand FAS2040 without disks back to live?

Re: Adding Nodes to Cluster With Used Shelves

Re: Expand the file Cifs aggr

Re: 7-mode Failover and Local Accounts

Re: convert_ucode, create_ucode-use

Re: Oncommand application shown volume size different

Re: Moving the aggregate...

Re: Configuring Direct FC connection to Server

Re: Adding Nodes to Cluster With Used Shelves

Re: Basic questions on Ontap upgrade 7 -mode- Can some help

Re: How to bring a second-hand FAS2040 without disks back to live?

Re: FAS2240 with DS14 MK4 disk shelves and cDOT.

Re: Using multiple LDAPs in single SVM

Re: cDOT - List all snapmirror relationships

Re: Moving the aggregate...

Re: Multiple cluster-management interfaces

Re: Multiple cluster-management interfaces

Re: Convert data disk as a spare disk

Re: CDOT 8.3 Mirror & Vault retention problems

Re: joining 2 clusters into 1

Re: joining 2 clusters into 1

Re: Snapmirror with same name and IP for destinatation

Re: Space release after volume delete

Re: Turning on/off updating of last access date for folders

Re: cDot Metro-Cluster Rename MetaData volume