toasters October 2006

toasters@lists.teaparty.net

30 participants
28 discussions

FAS960: backup speed vs simultaneous dumps...?
by Glascock, Donald 11 Oct '06

11 Oct '06

Hi, Toaster folks -- We're a small research group on our third Filer, which is a single FAS960 with five shelves on two cross-connected 2-Gb loops. We trigger flexVolume-level dumps of our one aggregate directly to 2-Gb-fibre-connected LTO-3 drives. When we dump two volumes simultaneously to two tape drives, the time taken to dump each volume increases by a third to a half again, compared to dumping each volume independently. For example, a stand-alone dump may take two hours, but if another dump of another volume to another drive is run simultaneously, then that two-hour dump becomes a three-hour dump. Adding a third simultaneous dump to a third drive causes an even greater increase in dump times. For diagnostic purposes, I've launched a conventional dump of one volume triggered via NDMP our backup software, waited for the dump to reach Phase-IV (the writing-data phase), and launched from the FAS960's console a dump of another volume to the FAS960's null device. Soon after this second dump starts, the write speed on the tape drive associated with the first dump falls off by quite a bit. This write speed picks up instantly when I kill the second dump. Since the second dump was only in Phase-I (inspecting files, but not writing them), and since this second dump was to the local null device anyway, I'm led to think that the performance issues lie somewhere within the Filer's head and Disk I/O subsystem. With one-at-a-time backups of our volumes, our full backup window is about eight hours (for about two terabytes). We were planning to buy two more shelves in early 2007. Should I expect a FAS960 to handle two or three simultaneous full dumps without a significant loss in dump speed? Are you getting nearly-linear performance out of your simultaneous dumps? Thanks for your time & have a great day! Don Glascock -- Donald S. Glascock Mailstop RO-SN-2-SPPDG Special Purpose Processor Development Group Mayo Foundation 4001 41st Street NW Rochester, MN 55901 Glascock.Donald(a)Mayo.EDU +1.507.538.5467 +1.507.284.9171 (fax) http://www.mayo.edu/sppdg/ "No matter where you go, there you are."

3 4

Statistics and Forecasting
by Carl Howell 11 Oct '06

11 Oct '06

What are you guys doing to provide storage statistics to your administrators for growth and budget forecasting? We run DFM, but I'm struggling to translate the statistics we get from it into something the administration can understand and act upon. Thanks in advance for the the help, Carl Howell Systems Engineer University of West Florida

2 1

Statistics and Forecasting
by Carl Howell 11 Oct '06

11 Oct '06

What are you guys doing to provide storage statistics to your administrators for growth and budget forecasting? We run DFM, but I'm struggling to translate the statistics we get from it into something the administration can understand and act upon. Thanks in advance for the help, Carl Howell Systems Engineer University of West Florida

2 1

Re: Problem renewing maintenance
by Chris Thompson 10 Oct '06

10 Oct '06

You may recall that back in June I wrote: > We're having a problem renewing maintenance cover on our F810. > I wonder whether anyone else has had a similar experience. > > It's being done through a secondary supplier who had better remain > nameless for the moment (just in case it's their fault rather than > NetApp's...). We originally brought 3 years of cover: > > "Next Business Day Parts Delivery" for failed disks (which has > worked fine, incidentally); > > "4 hours onsite service" for the base unit and disk shelf. > > Now we are being told that we can't have such a combination: it has > to be all NBD Parts or all 4hrs Onsite. This doesn't make any sense > to us. Disk failures are non-critical, and only the ... (ahem) very > differently abled ... need someone to come and swap a disk for them! > OTOH, base unit & shelf failures would be critical for us. > > Please tell me that NetApp haven't _really_ made it this difficult > to buy the obviously appropriate combination of cover! which got a number of helpful responses, and I then wrote: > Stephen C Woods <scw(a)seas.ucla.edu> wrote > > > Yes they have. We got nailed whwn we renewed our support last > > November. OUCH! > > Michael A Schipp <mschipp(a)asi.com.au> wrote: > > > Stephen is correct, Netapp did change it! However if you get > > you supplier to talk to Netapp they should be able to request a PRV from > > Netapp to renew the maintenance (Should is up to Netapp to approve - so > > may or may not happen). > > Thanks for this information, and I'll certainly try and get that last > suggestion acted on by our supplier. Well, our secondary suppliers (Majenta Solutions, for the record, and they are totally blameless in all this as far as I can see) did finally manage to get us the following deal 4 hour onsite cover for the base filer Next Busness Day parts delivery for the shelf and disks (it proved impossible to distinguish the shelf from the disks) for about the mean of the all-4hr and all-NBD quotes. So we went for that. Or so we thought... We have had two disk failures since then [*] and on each occasion NetApp shipped a replacement disc to the 4-hour schedule. So they must think that is what we paid for. I suppose it ought to be flattering that NetApp think we require a replacement disk that fast, but actually it's a pain in the posterior, as we aren't set up to accept deliveries 7x24. I suppose next year we'll say To Hell With It, and go for NBD for everything. [*] an unusually high rate in a 14-disc system, on past experience: one of them was just after a reboot upgrading ONTAP, the other a few hours after a scheduled outage of our Uninterruptible Power Supply ... no, don't ask, you don't want to know :-( -- Chris Thompson Email: cet1(a)cam.ac.uk

1 0

Quotes?
by Grey Friday 10 Oct '06

10 Oct '06

Has it been anyone else's experience that IBM is beating the pants off of other resellers? I'm talking by 50%? I'd like to hear from you if you find this the case as well.

5 5

RE: Removing a disk from an aggregate
by Fox, Adam 10 Oct '06

10 Oct '06

This is generally not possible. The only disks that can be removed from an aggregate is the dparity disk by changing the RAID type from RAID-DP to RAID-4. And you can only do this if your current RG is not over the max size for RAID-4. Sorry. -- Adam Fox NGS Tools Developer adamfox(a)netapp.com -----Original Message----- From: Alan Biren [mailto:ABiren@asisystem.com] Sent: Tuesday, October 10, 2006 12:37 PM To: toasters(a)mathworks.com Subject: Removing a disk from an aggregate Does anyone know of a way to permanently remove a disk from one a RAID Group in an Aggregate. One of my Administrators create an aggregate with raidsize of 14 disks and used all available disks to fill the raid group. There are now no spares in the system. The current aggregate already has a production volume on it so destroying the aggregate is really not an option. Thanks ************************************************************************ ***************************** Note: ASI System Integration Inc, is located at 48 West 37th Street, New York, NY 10018. The contents of this email and any attachments are confidential. It is intended for the named recipient(s) only. If you have received this email in error please notify the system manager or the sender immediately and do not disclose the contents to anyone or make copies. ** eSafe scanned this email for viruses, vandals or malicious content ** ************************************************************************ *****************************

1 0

RE: Estimating Aggregate IOPS
by Glenn Walker 05 Oct '06

05 Oct '06

Actually, That's a function of the ESH (6 shelves * 14 drives = 84 disks per loop). What I'm referring to is a bit different: as you ramp up the number of drives, the performance curve begins to increase - however, after you reach a certain number of drives, the sheer work behind laying the data out on them all (and retrieving it) begins to take it's toll and flatten the performance curve. (I would assume that contention on the loop would have something to do with this as well, not that it's saturated necessarily, but that there is a bit more latency as the disks take turns passing data). Glenn -----Original Message----- From: Suresh Rajagopalan [mailto:SRajagopalan@williamoneil.com] Sent: Thursday, October 05, 2006 1:32 PM To: Glenn Walker Cc: toasters(a)mathworks.com Subject: RE: Estimating Aggregate IOPS I see max drives per FC loop to be 84 on all filer models. Is this the limit you are alluding to? Thanks Suresh -----Original Message----- From: Glenn Walker [mailto:ggwalker@mindspring.com] Sent: Wednesday, October 04, 2006 3:57 PM To: Suresh Rajagopalan Cc: toasters(a)mathworks.com Subject: RE: Estimating Aggregate IOPS There is a ceiling in performance gain after a certain number of disks - it's well below 112, but I'm pretty sure it's above 56 (though not much above it). At that point, the 'curve' starts to go back down again. However, RAID-DP and RAID-4 (it's poor, homely cousin) are very fast and the filer has always handled the XORs very well for the parity calculations. A better question\way to position this might be what type of workload\performance you expect\require. While the disks are capable of about 120 IOPS for 10k and 180 IOPS for 15K disks, the latencies begin to jump at the end of the scale - if you're workload requires lower latencies, it doesn't much matter what the disk is capable of (Exchange is a very good example of this - 80 IOPS for 10k drives is about the max you'd want to go). Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Suresh Rajagopalan Sent: Wednesday, October 04, 2006 3:56 PM Cc: toasters(a)mathworks.com Subject: RE: Estimating Aggregate IOPS Let's assume 4k IOPS. Disks are currently 10k rpm. Are you saying that NTAP's implementation of RAID4 (or RAID-DP) gives a linear performance increase with spindles, (almost like RAID-0) and no penalty for the RAID4 or RAID-DP? That is, aggregate IOPS is pretty much (N * disk-iops) and no penalty for RAID4 or RAID-DP? Thanks Suresh -----Original Message----- From: Blake Golliher [mailto:thelastman@gmail.com] Sent: Wednesday, October 04, 2006 12:41 PM To: Suresh Rajagopalan Cc: toasters(a)mathworks.com Subject: Re: Estimating Aggregate IOPS How fast are the spindles, and what size is your iop? Generally, you can assume 110 per 10k RPM disk, and 180 for 15k RPM disk. This is also assuming a 10ms latency for each iop. And subtract 2 disks per raid group for raid dp (if doing writes, don't if you are doing pure reads). So with around 50 disks (after you subtract the raid dp overhead) you can expect around 5500 iops from that set of disks (assuming 10k rpm disks). For the 112 aggregate, you an expect 98 spindles and 10780 iops. I'm assuming 4k iops, and 10ms latency. Hope that helps, -Blake On 10/4/06, Suresh Rajagopalan <SRajagopalan(a)williamoneil.com> wrote: > Given a disk IOPS of 100, I'd like to estimate total aggregate IOPS for > the following cases: > > 1) 56 disks, 1 aggregate, RAID-DP size 16 > 2) 112 disks, 1 aggregate, RAID-DP size 16 > > I'm only interested in the total raw disk IOPS available in each case, > not including considering the filer head. For example, we know that > RAID-0 with 56 disks @100 would yield 5600 IOPS. > > I don't know how to do this calculation with Data ONTAP's implementation > of RAID4 or RAID-DP. > > Any assistance would help. > > Thanks > Suresh > > > > >

1 0

RE: Interesting FAS250 note on Tape connectivity
by Oliver Bassett 05 Oct '06

05 Oct '06

Glenn, Yes I have seen a couple of these now. I have set one up with direct connect to a tape library via optical FC using NDMP to control the backups. It isn't acutally optical LC until you put the GBIC in, so yeah it is a lot more flexible than before. As for the in the field replacement, I suspect you probably would end up with one of the ROHS units but it isn't a definate thing. If you are selecting the RMA part yourself select the one that has a part number ending in R5 as this is the ROHS kit. Regards Oliver -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Glenn Dekhayser Sent: Friday, 6 October 2006 8:07 a.m. To: toasters(a)mathworks.com Subject: Interesting FAS250 note on Tape connectivity From the NOW Forums: "FAS250 units ordered after 8-May-2006 are now fitted with optical LC (instead of copper HSSDC2) connector style. This change in connection with ROHS compliance." This was a major pain in the keester before, as there were only a few copper-FC tape devices available. Has anyone seen these newer units? I didn't see any announcements made on this, and the online system config guide still shows the copper HSSDC2 port, not fiber. Also, if there was a FAS250 in the field that needed a controller replacement, would Netapp ship the new RoHS compliant controller, or the older one? Glenn (the other one). The information contained in this email is privileged and confidential and intended for the addressee only. If you are not the intended recipient, you are asked to respect that confidentiality and not disclose, copy or make use of its contents. If received in error you are asked to destroy this email and contact the sender immediately. Your assistance is appreciated.

1 0

Interesting FAS250 note on Tape connectivity
by Glenn Dekhayser 05 Oct '06

05 Oct '06

From the NOW Forums: "FAS250 units ordered after 8-May-2006 are now fitted with optical LC (instead of copper HSSDC2) connector style. This change in connection with ROHS compliance." This was a major pain in the keester before, as there were only a few copper-FC tape devices available. Has anyone seen these newer units? I didn't see any announcements made on this, and the online system config guide still shows the copper HSSDC2 port, not fiber. Also, if there was a FAS250 in the field that needed a controller replacement, would Netapp ship the new RoHS compliant controller, or the older one? Glenn (the other one).

1 0

RE: Estimating Aggregate IOPS
by Suresh Rajagopalan 05 Oct '06

05 Oct '06

I see max drives per FC loop to be 84 on all filer models. Is this the limit you are alluding to? Thanks Suresh -----Original Message----- From: Glenn Walker [mailto:ggwalker@mindspring.com] Sent: Wednesday, October 04, 2006 3:57 PM To: Suresh Rajagopalan Cc: toasters(a)mathworks.com Subject: RE: Estimating Aggregate IOPS There is a ceiling in performance gain after a certain number of disks - it's well below 112, but I'm pretty sure it's above 56 (though not much above it). At that point, the 'curve' starts to go back down again. However, RAID-DP and RAID-4 (it's poor, homely cousin) are very fast and the filer has always handled the XORs very well for the parity calculations. A better question\way to position this might be what type of workload\performance you expect\require. While the disks are capable of about 120 IOPS for 10k and 180 IOPS for 15K disks, the latencies begin to jump at the end of the scale - if you're workload requires lower latencies, it doesn't much matter what the disk is capable of (Exchange is a very good example of this - 80 IOPS for 10k drives is about the max you'd want to go). Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Suresh Rajagopalan Sent: Wednesday, October 04, 2006 3:56 PM Cc: toasters(a)mathworks.com Subject: RE: Estimating Aggregate IOPS Let's assume 4k IOPS. Disks are currently 10k rpm. Are you saying that NTAP's implementation of RAID4 (or RAID-DP) gives a linear performance increase with spindles, (almost like RAID-0) and no penalty for the RAID4 or RAID-DP? That is, aggregate IOPS is pretty much (N * disk-iops) and no penalty for RAID4 or RAID-DP? Thanks Suresh -----Original Message----- From: Blake Golliher [mailto:thelastman@gmail.com] Sent: Wednesday, October 04, 2006 12:41 PM To: Suresh Rajagopalan Cc: toasters(a)mathworks.com Subject: Re: Estimating Aggregate IOPS How fast are the spindles, and what size is your iop? Generally, you can assume 110 per 10k RPM disk, and 180 for 15k RPM disk. This is also assuming a 10ms latency for each iop. And subtract 2 disks per raid group for raid dp (if doing writes, don't if you are doing pure reads). So with around 50 disks (after you subtract the raid dp overhead) you can expect around 5500 iops from that set of disks (assuming 10k rpm disks). For the 112 aggregate, you an expect 98 spindles and 10780 iops. I'm assuming 4k iops, and 10ms latency. Hope that helps, -Blake On 10/4/06, Suresh Rajagopalan <SRajagopalan(a)williamoneil.com> wrote: > Given a disk IOPS of 100, I'd like to estimate total aggregate IOPS for > the following cases: > > 1) 56 disks, 1 aggregate, RAID-DP size 16 > 2) 112 disks, 1 aggregate, RAID-DP size 16 > > I'm only interested in the total raw disk IOPS available in each case, > not including considering the filer head. For example, we know that > RAID-0 with 56 disks @100 would yield 5600 IOPS. > > I don't know how to do this calculation with Data ONTAP's implementation > of RAID4 or RAID-DP. > > Any assistance would help. > > Thanks > Suresh > > > > >

1 0

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 October 2006