About aborzenkov

aborzenkov · ‎2015-11-21

Node is controller. HA pair always has two controllers (nodes), in your case in the same chassis. To get transparent failover between two locations you need metro cluster, but metro cluster is not possible with FAS2xxx. Alternative is host-based mirroring (simply allocate storage from both HA pairs to VM and mirror on VM level).

aborzenkov · ‎2015-11-19

In 7-Mode licenses are kept in root volume, so if you restore it it should restore licenses as well.

aborzenkov · ‎2015-11-19

AFAIU the only way to do is to FlexClone destination and connect to some VM to get access to files. I suppose some software may automate it. It may be possible to file clone individual LUN, but in essence this is the same.

aborzenkov · ‎2015-11-19

You can mark aggr1 as root in maintenance mode (aggr options aggr1 root); it will create empty volume AUTOROOT and mark it as root. You will need to reconfigure filer from scratch though (including entering all licenses), or restore content of root volume from backup. It is rare that all disks fail at once, so it may give you chance to troubleshoot and recover your root aggregate though.

aborzenkov · ‎2015-11-19

Do you have 7-Mode or C-Mode?

aborzenkov · ‎2015-11-18

With LAN aggregate you depend on switch for load balancing. Cisco 3750 can distribute packets basing on MAC or IP, which means traffic between server and NetApp will always use single physical interface only. With dNFS you are better off to configure each physical link individually and relying on dNFS for load balancing and failover.

aborzenkov · ‎2015-11-17

Communication is between two IPs; once FQDN was resolved to IP it becomes irrelevant. If IP changed, host must again resolve FQDN to new IP and reestablish connection. Old connection will appear hung.

aborzenkov · ‎2015-11-17

fcp config or fcp show adapter You probably mean FAS 3250.

aborzenkov · ‎2015-11-16

Good question. My understanding is, "files" refers to /etc/passwd, and users added by useradmin are kept elsewhere (directly in registry). But I love to be proved correct (or wrong 🙂 )

aborzenkov · ‎2015-11-16

I do not say what you should do, it is up to you to decide; but you could do it, yes.

aborzenkov · ‎2015-11-16

Yes, you can do it. By default NetApp will leave last group incomplete, so simply set desired rg size when creating aggregate. It does not really matter in which order raid groups will be.

aborzenkov · ‎2015-11-15

All versions of Data ONTAP I am aware of supported hot shelf addition. Do you know version that did not?

aborzenkov · ‎2015-11-12

Oh, I missed this part. Sorry, I let someone else to chime in.

aborzenkov · ‎2015-11-12

I can't find the exact details of how fsecurity deals with existing DACLS thoug All information I have seen suggests fsecurity completely replaces existing DACL.

aborzenkov · ‎2015-11-11

You must mount the volume in SVM; you do it by creating a junction path. This path is the path you use in sharing it.

aborzenkov · ‎2015-11-11

Is user mapping configured for a domain D? If not, do you have default Unix user on filer? Otherwise there are deveral knowledge base articles that deal with it; just search for "trusted" on NetApp support site. You may want to check what is suggested there.

aborzenkov · ‎2015-11-11

You need to configure user mapping from Unix to Windows users. Without valid Windows user access is denied.

aborzenkov · ‎2015-11-10

It was mentioned recently that the latest Java requires larger key size (at least 1024 bits). Try to reconfigure SSL on filet specifying larger size.

aborzenkov · ‎2015-11-10

It should not disrupt connectivity on base interface. Yes, adding broadcast domain for new network is highly recommended. You do not really need IPspace normally, but it depends on your specific requirements (e.g. If you need different default route in new VLAN you probably need IP space).

aborzenkov · ‎2015-10-30

For CIFS access user mapping from Windows to Unix MUST succeed, even for access to folders with ntfs security style. In Windows-only environment it means, default Unix user MUST be defined as fallback. Which is pretty much confirmed by your disabling of Administrator-to-Unix mapping. Set default Unix user in properties, make sure this user is also defined in SVM.

aborzenkov · ‎2015-10-29

It probably is more useful to open new thread with clear subject instead of reviving years old catch all one ... Anyway You need online redo logs if you are creating *conistent* online backup without putting database in backup mode. That gives you crash-consistent image. You do not really need online redo logs if you put database in backup mode. In this case you need to replay archive logs past "end backup" anyway, so any online redo log saved as part of backup is outdated and of no use. Or may be he means to save online redo logs in addition to archived logs? That makes sense too as it gives you more recent data, thus less data is lost.

aborzenkov · ‎2015-10-29

If you have VSM, just snapmirror from old destination to new destination and then resync from source to new destination. You can also do incremental updates between old to new destinations to decrease cutover time as usual. vol move will probably work, but I have never used it for this; it has no real advantage over cascaded snapmirror. If you are using QSM, it is still possible, there is KB describing how to migrate QSM destination. In essence it is the same but there could be some subtle differences in handling; do not remember right now.

aborzenkov · ‎2015-10-27

For iSCSI it is better to use native host based multipathing across individual links; it is more flexible and gives you true load balancing. Otherwise multi is static and trusts you to configure partner correctly and LACP is using heartbeat to detect each link health. So in theory LACP can detect link failure even if physically interface is up; and it provides some protection against misconfiguration. They are the same as far as load distribution is concerned.

aborzenkov · ‎2015-10-27

What is "IP storage"? iSCSI?

aborzenkov · ‎2015-10-27

i can already boot_backup to 8.2.3 clustered "Panics on boot" does not fall to me into "can boot" category 🙂 i have started to wipe the filer now after booting to the 8.2.3C backup image, if its wiping the whole thing now in 8.2.3 then the panic will be gone and i can revert to 7 mode, correct ? I would be very surprised if it worked. You need to downgrade disk labels and for this you must perform downgrade from 8.3 to 8.2. Booting backup image is not downgrade.

Re: Hight availability with two FAS 2552

Re: root aggr disks failed and can't boot

Re: Restore from SNAPVAULT

Re: root aggr disks failed and can't boot

Re: root aggr disks failed and can't boot

Re: FC vs LACP (1gigabit members) performance

Re: NFS with vfiler and vfiler-dr

Re: WPN

Re: When to use security.admin.authentication internal or nsswitch?

Re: different number of disk in Raid Groups of the aggregate

Re: different number of disk in Raid Groups of the aggregate

Re: Adding a new SATA DS4246 in FAS2240 with DS2246 SAS

Re: FAS8020 new subnet, VLAN, ifgrp

Re: Setting NTFS permissions on OnTap shares

Re: CIFS - creating new mount points on the volume

Re: CIFS Access issue.

Re: UNIX permission to NTFS volumes in C-mode

Re: OnCommand System Manager recieves error 500

Re: FAS8020 new subnet, VLAN, ifgrp

Re: CIFS only Domain Admin has Access

Re: Snap Creator and Oracle RAC support

Re: How to migrate snapmirror destination volumes from FC to SAS aggregate

Re: Network Interface Group -- mode: Multiple versus LACP

Re: Network Interface Group -- mode: Multiple versus LACP

Re: C-DOT 8.3 Downgrade to 8.2.3