About aborzenkov

aborzenkov · ‎2012-06-30

3. Can the iGroups be made global between clustered FAS controllers. I am not sure about it. There is no requirement that igroups be identical between controllers. Different controllers may have LUNs presented to different hosts. Forcing identical igroups will create dummy entries that are not used at all. Or atleast have the ability to easily copy all the initators of an existing iGroup to another controller from withing the System Manager GUI? I agree, that would be useful. If LUN has to be presented to the same set of hosts it is really better to name igroup identically. OCSM could also logically merge both igroups to provide flat LUN ID space. Probably the most user friendly way would be asking user for LUN on controller and "host group" (to avoid confusion) and automatically creating/duplicating igroup on all controllers that contain LUNs presented to this "host group".

aborzenkov · ‎2012-06-29

Should be in /etc/log/stats/archive on current versions. But I apologize, I was wrong. While some part of file is indeed XML, actual data is binary. The hour stats are in /etc/log/cm_stats_hourly and are pure text.

aborzenkov · ‎2012-06-29

You are right. Quoting Active/Active guide: "Any dump commands directed to the failed node’s tape drives are executed on the takeover node’s tape drives." Good to know. I never can make myself to go and read through all the available manuals

aborzenkov · ‎2012-06-29

Do you mean that failover for NDMP backup to directly connected tape is possible at all? Do you have links to any description how it is setup?

aborzenkov · ‎2012-06-29

That is interesting question and I was not able to find definite answer so far nor examples how taken over partner can work with tape drive physically connected to a partner. So I simply assume that in takeover state backup won't work. This is not that big problem, as you usually do not run in this state for long time. In this case you do not need anything on partner. There should be no issues with backing up volumes via vfiler0. Scott likely has more experience

aborzenkov · ‎2012-06-29

Registry errors. Well ... Filesystem 1K-blocks Used Avail Capacity Mounted on localhost:0x80000000,0x69944c 180795672 180795672 0 100% /mroot root volume is indeed full. It may be possible to recover, but I would not trust myself guiding someone through it via forum.

aborzenkov · ‎2012-06-29

Do you have any user data on this filer? If not, reinitializing (^C for Boot Menu and select option 4) would be the simplest way. This has to be done when filers are not in takeover state. Make sure to have your licenses as you will need to reenter them again. Otherwise opening case with NetApp support is needed to make sure you will not accidentally lose data.

aborzenkov · ‎2012-06-29

32 to 64 does initiate conversion after snapmirror break. Of course, it is "32 bit volume on 64 bit aggregate" so conditions may not be exactly the same ...

aborzenkov · ‎2012-06-29

Please copy and paste exact console output after filer power on and until next reboot.

aborzenkov · ‎2012-06-29

Well ... if you are experienced programmer you may find it easier to use Data ONTAP SDK with Perl/Java/PowerShell/C# (and I guess there are are a couple more bindings). This at least avoids need to rely on uncommitted text representation. Performance counters are also collected by filer as hourly stats and per-second performance archives. They are just XML files after all and so processing could be automated as well. Format is not (publicly) documented, but can be guessed with enough motivation

aborzenkov · ‎2012-06-28

yes to both.

aborzenkov · ‎2012-06-28

You can map UUID to pretty disk name(s) using (e.g.) output of "storage show disk -v".

aborzenkov · ‎2012-06-28

No. You can copy snapshot content to volume on different aggregate using ndmpcopy or vol copy, but not snap restore it.

aborzenkov · ‎2012-06-28

Yes, that’s possible, exactly as you describe it. If you are going to remove single aggregate on FAS2020, you will need then to reinitialize it and do minimal setup, then after snapmirroring original root volume back mark it as root.

aborzenkov · ‎2012-06-27

It does not matter. Any interface that can reach partner RLM/SP/whatever would work technically. HW assist is entirely optional anyway, although I have seen it in action once.

aborzenkov · ‎2012-06-27

I believe it is actually using interface which corresponds to filer hostname by default. If my memory serves me right, hostname must be associated with interface, I got weird errors during boot if it was not the case.

aborzenkov · ‎2012-06-26

Host Utilities were never necessary for operation, they simply helped to automate host settings. During setup of Microsoft Cluster some disk IO related parameters may get reset (it was definitely the case in the past). So you have to check all settings documented in Host Utilities manual and correct them if changed. Host Utilities could as be run in Recover mode to reapply these settings; I do not remember if DSM also has an option to reapply settings post-installation.

aborzenkov · ‎2012-06-25

Was it enough to just boot until special menu? Because normally you cannot fully netboot into running Data ONTAP with 8.x; you stop in special menu. Did you do anything in special menu?

aborzenkov · ‎2012-06-25

Could you elaborate on how you fixed it (more detailed steps)? In particular, what "pointing it to the DOT8 upgrade image" means?

aborzenkov · ‎2012-06-24

Could someone explain how lost write protection works or point to any materials?

aborzenkov · ‎2012-06-24

I have 4 filers connected to two groups of switches (Cisco 3750E and Nexus 5010) with 1G and 10G ports. Filers are FAS3170 and FAS6240 with 8.0.2P6. I enabled cdpd on filers but got extremely weird results a) switches do not show 10G filer ports at all b) switches correctly show 1G filer ports connected to 3750 (there is no 1G ports on Nexus) c) filers do not actually show any port on connected Cisco, neither 1G nor 10G d) filers do show directly connected ports of each other (like ACP or one ports that were used for snapmirror seeding) I have another HA pair connected to Cisco 3750G with 1G interfaces only and exactly the same Data ONTAP and there it is mostly OK; at least filers do show connected switch ports. Both 3750 and Nexus correctly display connected ports of each other (with exception of Nexus management port, but I do not know if CDP is active there at all). Anyone has seen this before? May be there is open BURT already, in which case I will try to open case.

aborzenkov · ‎2012-06-23

Did you try “system power on” in SP? Otherwise if you are near filer, you could simply turn power off/on if they are in separate chassis, or pull controller out of chassis and plug it in if it is single chassis. I am curious - why did you need to do it? I do not remember a single case when it was necessary to power cycle it.

aborzenkov · ‎2012-06-20

CLI of course ☺

aborzenkov · ‎2012-06-20

It does not do this automatically, you have to explicitly say you want to convert when adding disks. Otherwise it just rejects command if size will exceed 16TiB.

aborzenkov · ‎2012-06-20

Some options are not visible until you set them. Just try to set it. Although I am not sure this option matters at all, because FAS250 does not support either active/active configuration or FCP.

Re: Feature Requests: Alias for WWPN in System Manager GUI

Re: Gathering Disk Stats - how?

Re: NDMP over FC

Re: NDMP over FC

Re: NDMP over FC

Re: FAS2040 stuck in loop, won't boot

Re: FAS2040 stuck in loop, won't boot

Re: Data Ontap 8.1 upgrade - RLW_Upgrading process in the background BEWARE!!

Re: FAS2040 stuck in loop, won't boot

Re: Gathering Disk Stats - how?

Re: Help!Can root volume snapmirror to another filer?

Re: Gathering Disk Stats - how?

Re: Snaprestore query

Re: Help!Can root volume snapmirror to another filer?

Re: How to clear the options setting of cf.hw_assist.partner.address

Re: How to clear the options setting of cf.hw_assist.partner.address

Re: Cluster Failover (CF -Takeover and CF -Giveback)

Re: Booting wrong version on cf giveback

Re: Booting wrong version on cf giveback

Re: Data Ontap 8.1 upgrade - RLW_Upgrading process in the background BEWARE!!

CDP support in presence of 10G interfaces?

Re: Filer does not turn on after reboot -s

Re: upgrade 32bit aggr to 64bit

Re: upgrade 32bit aggr to 64bit

Re: Oncommand System Manager for Data Ontap 7.0.6