toasters November 2010

toasters@lists.teaparty.net

22 participants
9 discussions

Hot adding disk shelves to N3600 A20 (FAS2050C)
by Ray Van Dolson 30 Nov '10

30 Nov '10

Pardon the IBM-terminology -- I have an N3600 A20 (ONTAP 7.3.1) and have acquired two EXN4000's to expand our storage capacity. The N3600 A20 is currently configured as Active/Active with the existing set of 20 SAS disks. I'm reading through the documentation on NOW[1] on hot-adding shelves and am running into a few questions: - Is the EXN4000 the same as the DiskShelf14mk4? - How do I know if I have ESH2/ESH4 modules or LRC or ESH modules? Also, the NOW documentation[1] refers to the Data ONTAP 7.3 Active/Active Configuration guide for hot-adding of shelves since we are using an Active/Active setup. I presume these are the insructions I need to use and not the normal hot-add instructions[1]. Anything else of note I should be aware of? Thanks, Ray (No sales inquiries please) [1] http://now.netapp.com/NOW/knowledge/docs/hardware/filer/html/ds14hwsv/appa5…

3 2

RE: How to design for iops?
by Kennedy, Jeffrey 29 Nov '10

29 Nov '10

We tried the flexshare route and it had an impact but only slight. I suspect it's because there is still delay for other requests when this group has filled NVRAM and all that has to flush to disk before it can service the comparatively miniscule "other stuff". And once that slight interrupt has happened the process repeats. But that's only a suspicion. In any case, we're fast coming to the conclusion that this particular group just needs to be isolated, they do not play well with others. So we'll review the tool's characteristics and determine if high end is needed or if mid-range is ok once isolated. As for a BOGF (Bolt-On-Go-FasterTM) Avere is a front runner for write/read acceleration, which is also a possibility for us with this particular group. Thanks. Jeff Kennedy Qualcomm, Incorporated QCT Engineering Compute 858-651-6592 From: Jan-Pieter Cornet [mailto:johnpc@xs4all.net] Sent: Thursday, November 25, 2010 2:59 PM To: Kennedy, Jeffrey Cc: NDMP List (toasters(a)mathworks.com) Subject: Re: How to design for iops? On 2010 Nov 23, at 21:06 , Kennedy, Jeffrey wrote: Let me first define iops in this case. Metadata operations with a larger percentage of writes than reads. I have a group that will regularly drive 60-70k metadata operations per second that consist of updates more than half the time. 10Gb and PAM cards will help with reads but the writes are the killer. Today they are on shared 6070's with both PAM and 10Gb. I'm using PA to see the "other ops" which is where I got the 60-70k number. It doesn't break out read v write for metadata but based on 'sysstat' I got the over-half-write info. When they ramp up everyone else on that filer feels it. Is there something that can be done to improve this other than isolating them? You could give them a lower priority using the 'priority' command (available in ontap 7.2+, known as 'FlexShare'). That way, other operations would have priority over these writes. That said, make sure you have the basics right: the aggregates where the write go should have enough space: at most 80% can be used preferably less. It should have enough spindles. Add more disks if your disk utilisation is high or CP take too long. Move everything else off of that aggregate. Writes mainly put a load on the NVRAM card. A somewhat debatable performance boost would be to remove the clustering and interconnect, making your entire NVRAM card available for writes, instead of taking up half of the NVRAM card for the cluster partner (of course, going from cluster to a stand-alone system will seriously impact your uptime and fault tolerance. Make sure you know what you do). Doing a headswap for a system with more nvram will also be beneficial (but not exactly low cost or easy to implement). I'm not aware of hardware addons that boost write performance, but do ask your netapp sales rep, there might be something available. -- Jan-Pieter Cornet <johnpc(a)xs4all.net<mailto:johnpc@xs4all.net>> Systeembeheer XS4ALL Internet bv Internet: www.xs4all.nl<http://www.xs4all.nl> Contact: www.xs4all.nl/contact<http://www.xs4all.nl/contact>

1 0

R: Is there's a way to remove and reassign a disk?
by Milazzo Giacomo 27 Nov '10

27 Nov '10

Thanks all for answers. They're quite what I was thinking about "physics" :) Also if I could use "fail/unfail" the raid group I operate is always the same...so that I could remove a disk from it but this last want back a disk :) And all this stuff for the customer is worried about the overall performances. He's got now three rg, the first with 20, the second with 26 and the last with 6 disks and he's concerned about performances gave by the smaller rg influencing all the aggregate... Da: Davin Milun [mailto:davin.milun@gmail.com] Inviato: sabato 27 novembre 2010 14:55 A: Milazzo Giacomo Cc: toasters(a)mathworks.com Oggetto: Re: Is there's a way to remove and reassign a disk? No - you're stuck with your raidgroup configuration, short of destroying the aggr. If it really was just a specific incorrect DISK that was used, and you wanted to use another, you could use "disk replace" to do that - but that won't let you change raidgroup layout. Davin. On Sat, Nov 27, 2010 at 5:17 AM, Milazzo Giacomo <G.Milazzo(a)sinergy.it<mailto:G.Milazzo@sinergy.it>> wrote: Hi all, In my experience I think is not possibile but a customer of mine has this issue. He wanted to extend the two raid group of his aggregate but once increased the raid group size he forgot to use the syntax: >aggr add -d x.y -g raidgroup name So the last raid group has been extended and another raid group has been created with only 5 disks... He asked me how to go back and my answer is: "Copy the data, destroy the aggr and recreate it..." :) But I remember that times ago a customer of mine used the "disk fail" command just to try it and (not sure on syntax) he was able to "unfail" the disk. So, what, for this case, if "disk fail" is used to fail the disk, rebuild its data on a spare, the unfail the failed, zeroed and manually assigned to the right raid group? Dott. Giacomo Milazzo [cid:image001.jpg@01CB8E48.29D11970] Technical Account Manager Sinergy SpA Filiale di Roma * 00198. viale Regina Margherita, 269 ' (+39) 3406001045 0644245272 7 (+39) 0226922048 * Giacomo.Milazzo(a)Sinergy.it

2 2

Is there's a way to remove and reassign a disk?
by Milazzo Giacomo 27 Nov '10

27 Nov '10

Hi all, In my experience I think is not possibile but a customer of mine has this issue. He wanted to extend the two raid group of his aggregate but once increased the raid group size he forgot to use the syntax: >aggr add -d x.y -g raidgroup name So the last raid group has been extended and another raid group has been created with only 5 disks... He asked me how to go back and my answer is: "Copy the data, destroy the aggr and recreate it..." :) But I remember that times ago a customer of mine used the "disk fail" command just to try it and (not sure on syntax) he was able to "unfail" the disk. So, what, for this case, if "disk fail" is used to fail the disk, rebuild its data on a spare, the unfail the failed, zeroed and manually assigned to the right raid group? Dott. Giacomo Milazzo [cid:image001.jpg@01CB8E24.49A29630] Technical Account Manager Sinergy SpA Filiale di Roma * 00198. viale Regina Margherita, 269 ' (+39) 3406001045 0644245272 7 (+39) 0226922048 * Giacomo.Milazzo(a)Sinergy.it

1 0

GBIC question
by David McWilliams 23 Nov '10

23 Nov '10

We purchased a new FAS2020 and we didn't order a GBIC at the same time, we now want to connect to a FCAL tape drive. Can I use a standard GBIC or do I have to get one from NetApp? Sláinte, David "Build a man a fire he'll be warm for a day, set a man on fire and he'll be warm for the rest of his life" <http://en.wikipedia.org/wiki/Discworld> - Terry Pratchett <http://en.wikipedia.org/wiki/Terry_Pratchett> Checkout my photos - http://www.panoramio.com/user/1113507

3 5

RE: NDMP backup of snapmirror target
by Kennedy, Jeffrey 23 Nov '10

23 Nov '10

I would be very interested to hear from Netapp contributors if NDMP backups of volume snapmirror's has changed. My education at the time was that volume level snapmirror's are RO for all blocks, which means inode pointers could not change since that meant a change to blocks *in* the volume. QSM/Snapvault has never suffered from this issue. Jeff Kennedy Qualcomm, Incorporated QCT Engineering Compute 858-651-6592 From: Jeremy Webber [mailto:Jeremy.Webber@al.com.au] Sent: Sunday, November 21, 2010 6:22 PM To: Kennedy, Jeffrey Cc: Jeremy Webber; Fred Grieco; toasters(a)mathworks.com Subject: Re: NDMP backup of snapmirror target Maybe there were limitations in earlier versions of OnTAP but I can confirm that we do tape backups, via tape, from a snapmirror destination without having to break the snapmirror relationship. A snapshot is automatically created in the snapmirror destination for the duration of the NDMP job. I can't recall if subsequent snapmirror updates fail during the backup and am not able to test this right now. OnTAP 7.3.2. -jeremy On 17/11/2010, at 6:11 AM, Kennedy, Jeffrey wrote: Things may have changed in OT since I last was concerned with tape backups but as I recall a snapmirrored *volume* can't take the backup snapshot because it's RO. So you point the backup software at the newest snapshot in the RO volume and back up that view. The only reason the NDMP backup makes a snapshot is to guarantee consistency. An existing snapshot already has that. Jeff Kennedy Qualcomm, Incorporated QCT Engineering Compute 858-651-6592 From: owner-toasters(a)mathworks.com<mailto:owner-toasters@mathworks.com> [mailto:owner-toasters@mathworks.com] On Behalf Of Fred Grieco Sent: Tuesday, November 16, 2010 6:26 AM To: toasters(a)mathworks.com<mailto:toasters@mathworks.com> Subject: NDMP backup of snapmirror target I am snapmirroring several volumes to a DR site, and I would like to do tape backups at that site. How should I do them? Can I use NDMP? It seems like I can't do straight NDMP because the replicated target is read-only. I am currently creating flexclones from snapshots existing in the volume (nightly.0...) but flexclone then locks this snapshot and snapmirror will fail to update ( I'm not sure if this happens at the next update or when the source system attempts to remove that snapshot and copy it over... I wasn't watching). What do you generally do to get snapmirrored volume data to tape? TIA, Fred -- Jeremy Webber Senior Systems Engineer Animal Logic Pty Ltd T: +61 2 9383 4837 F: +61 2 9383 4801

5 6

How to design for iops?
by Kennedy, Jeffrey 23 Nov '10

23 Nov '10

Let me first define iops in this case. Metadata operations with a larger percentage of writes than reads. I have a group that will regularly drive 60-70k metadata operations per second that consist of updates more than half the time. 10Gb and PAM cards will help with reads but the writes are the killer. Today they are on shared 6070's with both PAM and 10Gb. I'm using PA to see the "other ops" which is where I got the 60-70k number. It doesn't break out read v write for metadata but based on 'sysstat' I got the over-half-write info. When they ramp up everyone else on that filer feels it. Is there something that can be done to improve this other than isolating them? Thanks. Jeff Kennedy Qualcomm, Incorporated QCT Engineering Compute 858-651-6592

1 0

Measuring IOPS
by Davies, Matt 18 Nov '10

18 Nov '10

Hi all, Does anyone know if there is a way to calculate the number of IOPS a filer is handling via SNMP or some way to script a collection ? Thanks Matt ____________________________ Matt Davies Director of International IT Operations General Atlantic 23 Savile Row London W1S 2ET Tel: +44 207 484 3203 Fax: +44 207 484 2803 Mobile: +44 777 559 4265 ____________________________ ________________________________ _____________________________________________________________ This e-mail (including all attachments) is confidential and may be privileged. It is for the exclusive use of the addressee only. If you are not the addressee, you are hereby notified that any dissemination of this communication is strictly prohibited. If you have received this communication in error, please erase all copies of the message and its attachments and notify us immediately at help(a)generalatlantic.com . Thank You.

12 13

NDMP backup of snapmirror target
by Fred Grieco 16 Nov '10

16 Nov '10

I am snapmirroring several volumes to a DR site, and I would like to do tape backups at that site. How should I do them? Can I use NDMP? It seems like I can't do straight NDMP because the replicated target is read-only. I am currently creating flexclones from snapshots existing in the volume (nightly.0...) but flexclone then locks this snapshot and snapmirror will fail to update ( I'm not sure if this happens at the next update or when the source system attempts to remove that snapshot and copy it over... I wasn't watching). What do you generally do to get snapmirrored volume data to tape? TIA, Fred

4 3

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 November 2010