About aborzenkov

aborzenkov · ‎2018-11-21

Spare disk was not zeroed so ONTAP detects it as part of aggregate. Just destroy this foreign aggregate and zero new spare.

aborzenkov · ‎2018-11-07

Ah, right. If this is new shelf this is of course an option.

aborzenkov · ‎2018-11-07

@SpindleNinja wrote: Just do one at a time and you should be fine. Did it change recently? I reviewed available documentation and it says IOM12 upgrade procedure is a disruptive operation System and storage must be powered down to complete swap operation of IOM3/IOM6 to IOM12

aborzenkov · ‎2018-11-06

@SpindleNinja wrote: What you can do is replace the IOM6s in the DS4246 with IOM12s and you'll be good. Can it be done non-disruptively?

aborzenkov · ‎2018-11-02

@AndreUnterberg wrote: You will only get the whole space available when all disks of an aggregate are swapped. That's new to me. I was grown to believe that it is impossible to grow aggregate by replacing disks. Just to be sure, have you done it or this is your expectation?

aborzenkov · ‎2018-10-29

Only you know your infrastructure and only you can decide how disruptive it will be. You may be using management interfaces for NDMP or you may be having older Data ONTAP which was using node management LIFs for some control traffic associated with data access, so prolonged unavailability of LIF may impact data access as well, to name a few. There is always route via takeover/giveback, in this case you recable node offline while keeping data access to other node.

aborzenkov · ‎2018-10-27

Is it possible to do "partial permanent switchover" for one SVM only, moving primary workload to another site in MetroCluster?

aborzenkov · ‎2018-10-27

I evaluate possibility to migrate existing cluster to MetroCluster. To my best knowledge it is still impossible to convert existing cluster into MetroCluster non-distruptively - it requires data evacuation and wipe clean of existing cluster which is out of question. Assuming I build new MC and have space to accommodate old data - migrating data itself is more or less trivial. The challenge is to replicate SVMs together with all configurations. Unfortunately, SnapMirror SVM also is not supported for MetroCluster (although it is unclear what exactly is not supported - SVM DR between two clusters in MC? Or any SMV DR where one side is MC?) So what options are to automate rebuilding of SVM on new cluster? Doing it manually is rather error prone, it is easy to miss something.

aborzenkov · ‎2018-10-27

D) it can be done non disruptively as long as you can do it port by port. Migrate LIFs off port, reconnect port to new switches, migrate LIFs back. Honestly, this is the only sane way to do it. If switch migration is rip and replace, this is obviously downtime (unless your team is capable to perform replacement in less than a minute). IP change means inevitable downtime for each client to switch to new IP. You can resume SnapMirror without performing full sync. Also after IP change.

aborzenkov · ‎2018-10-24

@AndreUnterberg wrote: It's just for safety reasons. So your Snapshots cannot fill up your data space. Is it something new in C-Mode? In 7-Mode snap reserve did not restrict snapshot “size” in any way, and it was always possible to fill up all available space.

aborzenkov · ‎2018-10-24

No, you do not. There is no difference in behavior with or without snap reserve.

aborzenkov · ‎2018-10-19

Usual recommendations when some IOM fails to join OOB ACP try to replug LAN cables (plug out, wait couple of minutes, plug back in) try to disable and then enable ACP again try to reseat IOM in question All failed, case with NetApp support is in order.

aborzenkov · ‎2018-10-09

You may consider initial replication from A to C, then continuing from B to C. Note that it is possible to initialize snapmirror by simply shipping shelves (or small filer) if everything else is impossible.

aborzenkov · ‎2018-10-04

You have IB ACP and firmware version that does not support it. You need to switch to OOB ACP, update firmware and then switch back to IB if desired.

aborzenkov · ‎2018-10-03

Doing this across aggregates on different nodes would distribute load across them; depending on actual workload this could be quite beneficial. Doing this across aggregates on the same node probably is not going to have any noticable effect (at least, on Linux), although in extreme cases one may hit individual device limits (mostly queue depth) both on host and storage side, so adding more device may help.

aborzenkov · ‎2018-10-02

@wolfy wrote: another problem im having is that i had created a NFS partition that had been mounted onto vcenter but since i had updated the verion to 9.1 the mpounted datastore is now (inactive) and cant unmount cuz it thinks it is still being used. The described procedure is destructive. All existing data on disks is lost. Your NFS partition you created in 7-Mode no more exists. You need to setup filer completely new after having installed 9.1 and having created cluster.

aborzenkov · ‎2018-10-02

I'm pretty sure that on FAS2240 set-defaults does not include bootarg.init.boot_clustered. Is it not needed with ONTAP 9.x anymore?

aborzenkov · ‎2018-10-01

There is no point in jumping through 8.3, any cDOT version is installed identically. Also following this KB article will create traditional root aggregate wasting quite some space; if going to cDOT, one should really start with ADP.

aborzenkov · ‎2018-09-30

It may be possible to force RDB resync; see https://library.netapp.com/ecmdocs/ECMP1196798/html/GUID-379190AA-72D7-4CC4-9BAF-DC509DA06C30.html I’d still open case with NetApp support to investigate.

aborzenkov · ‎2018-09-29

Takeover, swap cables, giveback.

aborzenkov · ‎2018-09-26

How can I import ASUP from both nodes as part of the same data collection job? I can import two ASUP separately, but then when I try e.g. to display stack diagram both say "No stacks to display" for partner node. Nor does it show partner nodes in visualization.

aborzenkov · ‎2018-09-25

Is ACP active? Paste output of "storage shelf acp show -instance"

aborzenkov · ‎2018-09-25

e0c and e0d are LAN ports, you probably mean 0c and 0d. Yes, two other ports can be used either as LAN or FC. You will need to use matching SFP and possible reconfigure port mode (switch between LAN and FC).

aborzenkov · ‎2018-09-09

It sounds like you have only root aggregate. In this case the most simple is to reinstall from scratch with both controllers. You will need to change ha mode in maintenance boot before doing it.

aborzenkov · ‎2018-09-08

FAS2620 is not supported by 7-Mode at all.

Re: Insufficient spare disks

Re: Adding a DS4246 to a FAS2720 with DS212C

Re: Adding a DS4246 to a FAS2720 with DS212C

Re: Adding a DS4246 to a FAS2720 with DS212C

Re: Disk replacement for a FAS2554

Re: Network switch change from Juniper to Cisco

Switching primary site for SVM in MetroCluster?

Replicating SVM configuration onto MetroCluster?

Re: Network switch change from Juniper to Cisco

Re: SnapShot behavior with or without SnapReserve enabled

Re: SnapShot behavior with or without SnapReserve enabled

Re: Upgrading IOM6 Firmware Before Moving Shelf

Re: 7-Mode cascaded SnapMirror initialization

Re: Upgrading IOM6 Firmware Before Moving Shelf

Re: Load balance between two aggregates

Re: Clean installing cDOT 9.1.x onto our FAS2240

Re: Clean installing cDOT 9.1.x onto our FAS2240

Re: Clean installing cDOT 9.1.x onto our FAS2240

Re: vifmgr and mgmt offline

Re: FAS2650 miniSAS Cable Connections - Need to Swap Ports

Config Advisor 5 - import HA pair ASUP for analysis?

Re: Upgrading IOM6 Firmware Before Moving Shelf

Re: FAS2552 ports

Re: FAS 2620 - Adding new controller to existing single controller chassis

Re: FAS 2620 - Adding new controller to existing single controller chassis