toasters December 2006

toasters@lists.teaparty.net

50 participants
40 discussions

V-Series with NDMP backup
by Ammar Zolkipli 20 Dec '06

20 Dec '06

Gents, Let me first explain the overall layout. I've got a D2D2T design. Both primary and secondary filers are v-series, and snapvaulted. We have 16TB of CIFS home dir data volumes backing up to tape via NDMP. And the tape backup is on LTO-3 running at sustained throughput of 20mbps, (F**king slow, I know). Does anybody have experience how to speed up NDMP backup? At this point, the LTO-3 is shoe-shining as the minimum sustained throughput on that drive is 40mbps. And also, does anybody have any experience with NDMP to tell us why this is happening? Thanks

5 8

RE: Aggregrate snap reserve (again)
by Glenn Walker 19 Dec '06

19 Dec '06

I'm pretty sure that the aggr snap reserve is not needed for core dump - this is written to the 10% overhead space (or to spare drives, depending on how you've configured your filer), then when the filer comes back up and the 'savecore' is performed, it is written to the /etc/crash directory. As for the aggr copy command, that is news to me (you see how often I've used it). The one thing you did overlook: restoring an aggregate - not individual volume, mind you. This is the same concept as vol snap restore. Glenn ________________________________ From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Michael Schipp Sent: Monday, December 18, 2006 5:54 PM To: toasters(a)mathworks.com Subject: Aggregrate snap reserve (again) Hi all, We have had a lot of talk since DOT 7 of this but just to clear it up some (I hope). aggregate snapshots are required if you use (default 5%); * RAID SyncMirror and/or MetroCluster configurations. * Aggr copy command * Core dump IS the core dump in on aggr0 snap space only? So if a Filer/FAS has two aggregates and not using or needing SyncMirror and the aggrcopy command is it safe to turn off the aggregate snap reserve for aggr1 (still leaving snap reserve on for aggr0 for core dumps)? NetApp - are aggregates snap reserve used for any other purpose? Thanks Michael

2 1

RE: Aggregrate snap reserve (again)
by Learmonth, Peter 18 Dec '06

18 Dec '06

Core dump does not go into aggr snap reserve. During panic, core is dumped into a reserved area (the space chopped off by rightsizing) on one or more non-broken disks (outside of WAFL, since WAFL code runs in memory, which we might no longer trust depending on what went wrong and we're dumping it). During reboot, savecore (see the man pages and System Admin Guide) moves the core from reserved area to /etc/crash, where it can be picked up by the user or other process. I could be wrong about which disks it uses, but I'm sure it doesn't use aggr snap reserve. Peter ________________________________ From: Michael Schipp [mailto:mschipp@asi.com.au] Sent: Monday, December 18, 2006 2:54 PM To: toasters(a)mathworks.com Subject: Aggregrate snap reserve (again) Hi all, We have had a lot of talk since DOT 7 of this but just to clear it up some (I hope). aggregate snapshots are required if you use (default 5%); * RAID SyncMirror and/or MetroCluster configurations. * Aggr copy command * Core dump IS the core dump in on aggr0 snap space only? So if a Filer/FAS has two aggregates and not using or needing SyncMirror and the aggrcopy command is it safe to turn off the aggregate snap reserve for aggr1 (still leaving snap reserve on for aggr0 for core dumps)? NetApp - are aggregates snap reserve used for any other purpose? Thanks Michael

1 0

Aggregrate snap reserve (again)
by Michael Schipp 18 Dec '06

18 Dec '06

Hi all, We have had a lot of talk since DOT 7 of this but just to clear it up some (I hope). aggregate snapshots are required if you use (default 5%); * RAID SyncMirror and/or MetroCluster configurations. * Aggr copy command * Core dump IS the core dump in on aggr0 snap space only? So if a Filer/FAS has two aggregates and not using or needing SyncMirror and the aggrcopy command is it safe to turn off the aggregate snap reserve for aggr1 (still leaving snap reserve on for aggr0 for core dumps)? NetApp - are aggregates snap reserve used for any other purpose? Thanks Michael

1 0

RE: CF Disk Configuration Question
by Learmonth, Peter 18 Dec '06

18 Dec '06

This is completely supported. I would suggest putting loopback plugs on the unused ports. This was best practice on FAS900, but I'm not sure about FAS3000. The loopback plugs slightly improve takeover and reboot times by not having the filer wait for timeout on an open loop. LC loopback is X6521-R6 http://now.netapp.com/eservice/partDetails.do?partNumber=X6521-R6&produc tId=60420&rohsCompliant=R6 I know it says FAS270 on there, but it is used on any LC connector where loopback is needed. You can even use it to test Ethernet ports, although there is no value in terminating unused Eth ports. Enjoy! Peter -----Original Message----- From: Stephen C. Losen [mailto:scl@sasha.acc.virginia.edu] Sent: Monday, December 18, 2006 4:33 AM To: toasters(a)mathworks.com Subject: CF Disk Configuration Question We have a 3050c clustered pair where both filers have 4 shelves of 144G FC disks each. We want to add two shelves of 500G SATA drives to one filer. We are not allowed to mix FC and SATA on the same FC Loop, so the two new shelves must be on a separate loop. Since this is a CF configuration, we will connect these shelves to both filers. The filer that owns the shelves will connect using the "A" port of its adapter while the partner filer will use its "B" port. But since we are not adding any shelves to the partner filer, this leaves us with a vacant "B" port on the filer that owns the shelves and a vacant "A" port on the partner. Does anyone know if this is a problem? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

RE: CF Disk Configuration Question
by Holland, William L 18 Dec '06

18 Dec '06

Worse case, put a fiber loopback plug in any vacant fiber ports. -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Willeke, Jochen Sent: Monday, December 18, 2006 8:59 AM To: Stephen C. Losen; toasters(a)mathworks.com Subject: RE: CF Disk Configuration Question Hi, we never run this configuration over a long time, but during cluster updates we had this configuration several times, even it was only for a couple of minutes. From the technical point of view it works perfectly. But i do not know if it is an supported scenario from netapp. Best Regards Jochen -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Stephen C. Losen Sent: Monday, December 18, 2006 1:33 PM To: toasters(a)mathworks.com Subject: CF Disk Configuration Question We have a 3050c clustered pair where both filers have 4 shelves of 144G FC disks each. We want to add two shelves of 500G SATA drives to one filer. We are not allowed to mix FC and SATA on the same FC Loop, so the two new shelves must be on a separate loop. Since this is a CF configuration, we will connect these shelves to both filers. The filer that owns the shelves will connect using the "A" port of its adapter while the partner filer will use its "B" port. But since we are not adding any shelves to the partner filer, this leaves us with a vacant "B" port on the filer that owns the shelves and a vacant "A" port on the partner. Does anyone know if this is a problem? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

CF Disk Configuration Question
by Stephen C. Losen 18 Dec '06

18 Dec '06

We have a 3050c clustered pair where both filers have 4 shelves of 144G FC disks each. We want to add two shelves of 500G SATA drives to one filer. We are not allowed to mix FC and SATA on the same FC Loop, so the two new shelves must be on a separate loop. Since this is a CF configuration, we will connect these shelves to both filers. The filer that owns the shelves will connect using the "A" port of its adapter while the partner filer will use its "B" port. But since we are not adding any shelves to the partner filer, this leaves us with a vacant "B" port on the filer that owns the shelves and a vacant "A" port on the partner. Does anyone know if this is a problem? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

2 1

RE: iSCSI Cluster help
by Glenn Walker 16 Dec '06

16 Dec '06

Yeah - the default SCSI timeout in windows is 10 seconds - that should be increased to at least 180 seconds as a failback could take up to 3 minutes. If you have snapdrive, it does this for you automatically. FYI - leaving it at only 10 seconds could be REALLY bad if you have a performance hungry application like Exchange (verification plus normal operation could create havoc with undersized spindle count). Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of rob-7704(a)austin.rr.com Sent: Friday, December 15, 2006 7:31 PM To: toasters(a)mathworks.com Subject: iSCSI Cluster help Hey you iSCSI wizards, We are just getting into iSCSI for winders and have the following scenario happen today: We had a failover on a cluster member that host test/infrastructure systems today, due to a faulty ESH2 (that was triggered by a background FW update, but that is a different discussion). This filer also has an iSCSI volume on a test window server. During a failover, it reported errors, but the iSCSI reconnected. However, during a giveback (after NetApp replaced the faulty ESH2), the iSCSI connection gave errors and lost the connection altogether. We had to reboot that test windows box to recover the iSCSI volume. We are working with NetApp to learn more about this. But just from browsing on the NOW site about iSCSI, there are recommendations about increasing the timeout values from the default 10 sec to at least 60 sec, and other registry tunings for iSCSI to make it more robust. http://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs16249 http://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs17115 Anybody been there and done this or other tunings with success ? TIA -Rob

1 0

iSCSI Cluster help
by rob-7704＠austin.rr.com 16 Dec '06

16 Dec '06

Hey you iSCSI wizards, We are just getting into iSCSI for winders and have the following scenario happen today: We had a failover on a cluster member that host test/infrastructure systems today, due to a faulty ESH2 (that was triggered by a background FW update, but that is a different discussion). This filer also has an iSCSI volume on a test window server. During a failover, it reported errors, but the iSCSI reconnected. However, during a giveback (after NetApp replaced the faulty ESH2), the iSCSI connection gave errors and lost the connection altogether. We had to reboot that test windows box to recover the iSCSI volume. We are working with NetApp to learn more about this. But just from browsing on the NOW site about iSCSI, there are recommendations about increasing the timeout values from the default 10 sec to at least 60 sec, and other registry tunings for iSCSI to make it more robust. http://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs16249 http://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs17115 Anybody been there and done this or other tunings with success ? TIA -Rob

1 0

Re:AFS
by Rohit 15 Dec '06

15 Dec '06

Hi Mike GX maybe the AFS flavor from NetApp. http://blogs.netapp.com/dave/TechTalk/?permalink=ONTAP-GX-151-Past-and-Futu… On the other hand, if you're looking for something _free_, try http://www.openafs.org/. AFS is killer replacement for NFS if you have lots of many small files that get constantly updated. Of course, not to mention, doing away with stale file handles as well. regards rohit Sphar, Mike wrote: > Just throwing a question out there curious to hear people's thoughts or > experiences. Every time I end up dealing with hundreds of stale file > handles because of a server move/change I become increasingly annoyed by > the stateless nature of NFS and think to myself "Maybe this time I'll > finally start seriously looking at AFS." > > Other than lots of other ways an AFS deployment could be complicated, I > wonder how if at all a Netapp can be part of an AFS deployment? > > Also feel free to tell me how using AFS is crazy in general and I should > just accept my stale file handles. > ---------------------------------------------------------------------- Online Criminal Justice Programs Criminal Justice careers are booming. Education-Advancement offers... http://tagline.bidsystem.com/fc/BgLEQfJAsToxV9QXVIZNjpNvgdu5bRXkSF30/

5 4

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 December 2006