toasters April 2007

toasters@lists.teaparty.net

68 participants
51 discussions

Netapp and VMware... It's all so confusing!
by Timothy Hollingworth 01 Apr '10

01 Apr '10

So, say for example, I want to virtualize an Exchange server. Lets also say that I want to manage it via Snapdrive and SME. Lets also say, just for the heck of it, that I want both the new front-engined Acura NSX as well as the Porsche RC3. The cars aside, I would have to run iSCSI from within the VM, install and use Snapdrive from within the VM to provision LUNs, install Exchange and then use SME? It seems kind of silly to use iSCSI from within a VM if the ESX box were fiber attached to the Filer, just to be able to use Snapdrive and SME, no? I would love to be able to use the features of SD and SME within a completely virtual environment and am wondering what some real world implementations other VMware/Netapp customers are running. Do you run storage for the databases and logs in VMDK's and manage snapshots and backups from a VMware virtual machine perspective or do you manage them as iSCSI LUNs and use SD and SME? Or am I completely blind and missing something? Timothy L. Hollingworth Sr. Network Engineer, ePlus Technology Inc. 678.462.6698 (cell) AIM: HollingworthTim YIM: t_holling

11 15

autosupport stops functioning
by Leeds, Daniel 27 Feb '08

27 Feb '08

anyone else have this happen? suddenly one of our filers stops generating autosupport messages. both autosupport.enable and autosupport.support.enable are on. if i generate a test email with the autosupport.doit option nothing happens. not even a failed message, the console just sits there as if i never generated one. needless to say neither netapp or our email server gets anything from this filer and nothing hits the console about generating any autosupport emails.

8 12

RE: Data migration
by Suresh Rajagopalan 27 Apr '07

27 Apr '07

I did try clearing the snapmirror snapshot but that did not work for me. Setting snapmirror.enable to be off worked. Suresh _____ From: Leeds, Daniel [mailto:dleeds@edmunds.com] Sent: Friday, April 27, 2007 1:06 PM To: Kleinbrahm, Robert; Suresh Rajagopalan; toasters(a)mathworks.com Subject: RE: Data migration or you can clear the snapmirror snapshot within the volume to remove the relationship. -- Daniel Leeds Senior Systems Administrator Edmunds.com -----Original Message----- From: Kleinbrahm, Robert [mailto:rkleinbrahm@firstrepublic.com] Sent: Fri 4/27/2007 1:03 PM To: Suresh Rajagopalan; Leeds, Daniel; toasters(a)mathworks.com Subject: RE: Data migration I had the same thing happen to me. Solution is to turn snapmirror off on both the source and destination before mounting the lun. Netapps response on this. 1. Turn off SnapMirror on source filers 2. Turn off SanpMirror on destination filers. Snapmirror no longer detects a previously SnapMirrored relationship. When connecting to that LUN, the mounting process takes very little time. This worked for me and is quite fast. Bob K. ________________________________ From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Suresh Rajagopalan Sent: Friday, April 27, 2007 12:05 PM To: Leeds, Daniel; toasters(a)mathworks.com Subject: RE: Data migration OK I tried these steps with a small 20g volume. Unmount the source lun, do final snapmirror update, break mirror, set fs_size_fixed on. Now When I try to connect the new lun via snapdrive on the destination, I get (after the prompt for SFSR) the following error message: Unable to connect disk Failure in checking policies Error: unable to find a snapshot containing a consistent virtual disk to restore This happens on both snapdrive 4.1 and 3.1.1R1. What am I missing? Thanks Suresh ________________________________ From: Leeds, Daniel [mailto:dleeds@edmunds.com] Sent: Thursday, April 26, 2007 3:26 PM To: Suresh Rajagopalan; toasters(a)mathworks.com Subject: RE: Data migration snapmirror is what you want. you can set it up anytime prior to the migration and let it run on a schedule or manually to build the initial baseline. after that you just do a snapmirror update to keep it in sync. we have done this so many times it's second nature. quite simply: 1) create target volume and initialize snapmirror relationship to source while lun is live and serving data 2) setup maintenance window and take filesystem/lun offline 3) perform final snapmirror update 4) break snapmirror relationship and bring new lun/volume online 5) mount new lun/filesystem and serve up data snapmirror is a really handy product :) -- Daniel Leeds Senior Systems Administrator Edmunds.com -----Original Message----- From: owner-toasters(a)mathworks.com on behalf of Suresh Rajagopalan Sent: Thu 4/26/2007 3:02 PM To: toasters(a)mathworks.com Subject: Data migration I need to migrate a 1.4Tb lun from one filer to another with minimal downtime (< 30 mins). So clearly some copying of data has to be done prior to the actual move. Would snapmirror or flexclone be the most appropriate for this? I'm not sure how I would get a consistent copy of the lun over to the other filer. This is not a database, the lun contains flat files on NTFS. Thanks Suresh The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. First Republic Bank and its subsidiaries do not take responsibility for, or accept time-sensitive instructions sent by email including orders, funds transfer instructions or stop payments on checks. All instructions of this nature must be handled by direct communication, not email. We reserve the right to monitor and review the content of all email communications sent or received. Emails sent to or from this address may be stored in accordance with regulatory requirements.

1 0

RE: Data migration
by Suresh Rajagopalan 27 Apr '07

27 Apr '07

OK I tried these steps with a small 20g volume. Unmount the source lun, do final snapmirror update, break mirror, set fs_size_fixed on. Now When I try to connect the new lun via snapdrive on the destination, I get (after the prompt for SFSR) the following error message: Unable to connect disk Failure in checking policies Error: unable to find a snapshot containing a consistent virtual disk to restore This happens on both snapdrive 4.1 and 3.1.1R1. What am I missing? Thanks Suresh _____ From: Leeds, Daniel [mailto:dleeds@edmunds.com] Sent: Thursday, April 26, 2007 3:26 PM To: Suresh Rajagopalan; toasters(a)mathworks.com Subject: RE: Data migration snapmirror is what you want. you can set it up anytime prior to the migration and let it run on a schedule or manually to build the initial baseline. after that you just do a snapmirror update to keep it in sync. we have done this so many times it's second nature. quite simply: 1) create target volume and initialize snapmirror relationship to source while lun is live and serving data 2) setup maintenance window and take filesystem/lun offline 3) perform final snapmirror update 4) break snapmirror relationship and bring new lun/volume online 5) mount new lun/filesystem and serve up data snapmirror is a really handy product :) -- Daniel Leeds Senior Systems Administrator Edmunds.com -----Original Message----- From: owner-toasters(a)mathworks.com on behalf of Suresh Rajagopalan Sent: Thu 4/26/2007 3:02 PM To: toasters(a)mathworks.com Subject: Data migration I need to migrate a 1.4Tb lun from one filer to another with minimal downtime (< 30 mins). So clearly some copying of data has to be done prior to the actual move. Would snapmirror or flexclone be the most appropriate for this? I'm not sure how I would get a consistent copy of the lun over to the other filer. This is not a database, the lun contains flat files on NTFS. Thanks Suresh

3 2

Data migration
by Suresh Rajagopalan 27 Apr '07

27 Apr '07

I need to migrate a 1.4Tb lun from one filer to another with minimal downtime (< 30 mins). So clearly some copying of data has to be done prior to the actual move. Would snapmirror or flexclone be the most appropriate for this? I'm not sure how I would get a consistent copy of the lun over to the other filer. This is not a database, the lun contains flat files on NTFS. Thanks Suresh

8 8

Hot-remove loop B
by Suresh Rajagopalan 26 Apr '07

26 Apr '07

We are running a 3020 cluster and need to do maintenance on one side of the cluster (filer head and disk shelves). Can we unplug loop B from the other cluster member while it is running? We only need it to service its loop A during the maintenance window. Thanks Suresh

2 1

Hot add first shelf to loop?
by Stephen C. Losen 26 Apr '07

26 Apr '07

We are going to start a new FC disk loop. The new shelf is DS14mk2 with ESH2 modules and we will connect it to a X2050B-R5 adapter. This is a FAS960c clustered pair running DOT 7.1.1.1 so we will connect the shelf to both heads. We want to do the same thing with a second shelf on the other head. Can this be done without taking down the filers? I checked NOW and only found instructions for hot adding a shelf to an existing loop, A little more info... We have a downtime scheduled to install the new FC adapters in the heads. We have two new shelves in a new cabinet for which we do not have power yet. If the circuits have been run before the downtime, then we are OK. If not, then we will install the adapters in the heads and connect the shelves later when we have power, but we want to avoid a second downtime for that. If we must power down the head to add the first shelf, then we can't use cf takeover and giveback because the shelf is connected to both heads. Another possibility would be to run extension cords to the new shelves and when we get power to the cabinet, switch the power plugs one at a time so the shelves don't lose power. The netapp cabinet is powered by 220V circuits, and I suppose the outlets in the cabinet are also 220V. Can we run extension cords from 110V circuits to power the shelves temporarily? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

5 6

Phantom LUN
by Theo Dijkstra 26 Apr '07

26 Apr '07

Hi there, I got a phantom lun showing up in my messages file when starting the iSCSI service: -------- [root@blue snapdrive]# sanlun lun show no filer LUNs available -------- snap1> lun show snap1> version NetApp Release 7.2.1: Sun Dec 10 00:23:08 PST 2006 -------- iscsi service start: Apr 26 09:22:24 blue iscsi: iscsi config check succeeded Apr 26 09:22:24 blue kernel: iscsi-sfnet: Loading iscsi_sfnet version 4:0.1.11-3 Apr 26 09:22:24 blue kernel: iscsi-sfnet: Control device major number 254 Apr 26 09:22:24 blue iscsi: Loading iscsi driver: succeeded Apr 26 09:22:29 blue iscsid[4950]: version 4:0.1.11-3 variant (02-May-2006) Apr 26 09:22:29 blue iscsi: iscsid startup succeeded Apr 26 09:22:29 blue iscsid[4954]: Connected to Discovery Address 192.168.1.81 Apr 26 09:22:29 blue kernel: iscsi-sfnet:host4: Session established Apr 26 09:22:29 blue kernel: scsi4 : SFNet iSCSI driver Apr 26 09:22:29 blue kernel: scsi: unknown device type 31 Apr 26 09:22:29 blue kernel: Vendor: NETAPP Model: LUN Rev: 0.2 Apr 26 09:22:29 blue kernel: Type: Unknown ANSI SCSI revision: 04 Apr 26 09:24:59 blue ntpd[3373]: kernel time sync enabled 0001 --- Where does it get it from? This happened after a crash of the linux box while I was testing some multipathing. This in hindering other operations like mounting new luns etc. -Theo

2 3

network troubleshooting
by Pazder, Steve B 25 Apr '07

25 Apr '07

Hello World! Our Network admin has been working with Cisco to troubleshoot some issues (limited to 1 switch on 1 particular floor). While our Filers are not involved in this issue, fingers are generally pointed in that direction just because they exist....on the network. Our network admin is looking for information from the Filers that would be akin to a TCP dump. Ifstat, netstat and statit output are not providing the information being looked for (handshake info, etc). I haven't run across this particular issue before and don't recall ever seeing it presented to Toasters. Does anyone know of a way to get the actual communication dumps (again, a TCP dump) from a Filer? Thanks! Steve

9 10

NetApp ESX Host Utilities?
by John Sellens 24 Apr '07

24 Apr '07

There was a recent mention of "NetApp ESX Host Utilities" on the list, and (apparently) a config_mpath command included therein. I've looked on the net and on now.netapp.com and I'm having no luck in locating these tools. If someone could provide a pointer, I would be much obliged. Thanks! John jsellens(a)generalconcepts.com

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters April 2007