toasters

toasters@lists.teaparty.net

3 participants
13514 discussions

Start a nNew thread

toasters
by TomArlington 30 Dec '99

30 Dec '99

toasters

1 0

RE: "Wac"-ing the WAFL?
by Mohler, Jeff 30 Dec '99

30 Dec '99

Yes, but in the spirit of keeping the post simple for our new user, I didnt bring in complexities of how a F720 and lower work, or clustering. So yes, my post is somewhat inaccurate outside the scope of the reply itself. :-----Original Message----- :From: tkaczma(a)gryf.net [mailto:tkaczma@gryf.net] :Sent: Thursday, December 30, 1999 8:51 AM :Cc: toasters(a)mathworks.com :Subject: RE: "Wac"-ing the WAFL? : : : :On Wed, 29 Dec 1999, Mohler, Jeff wrote: : :> When a series of writes comes into the box, it is buffered :*im making this :> simple*, and it flushed out to disk (on a 760) ever 10 :seconds, of when 16Mb of :> data is in the buffer (NVRAM). : :This is inaccurate, I believe in clustered machines with 32MB :NVRAM it is :8MB. I think the general rule is half of NVRAM dedicated to the :particular filer. : :Tom :

1 0

RE: "Wac"-ing the WAFL?
by Mohler, Jeff 30 Dec '99

30 Dec '99

:RAID 0. I know, I know, it sounds almost suicidal but as I :said speed is more :important in this instance than recovery. We'll be testing :this 760 against :an Auspex, which can be configured to use RAID 0, so I was :wondering if the :NetApps could also do this. --- Actually, the way RAID4 is used in conjunction with WAFL, really looks and feels like RAID0, just with a parity write at the end. WAFL removes the problem on a UFS type filesystem of having that single parity drive just thrash to tiny bits, taking the stripe performance with it. When a series of writes comes into the box, it is buffered *im making this simple*, and it flushed out to disk (on a 760) ever 10 seconds, of when 16Mb of data is in the buffer (NVRAM). A single stripe (or series of stripes) is then written in a _very_ optimized manner using the WAFL filesystem, and a single parity calculation/write is created per stripe. I direct you to a document Mr. Hitz wrote at http://www.netapp.com/tech_library/3001.html for more details. I have every confidence you will enjoy our performance in your testing.

3 2

volcopy and snapmirror performance
by kevin graham 30 Dec '99

30 Dec '99

We're looking at moving 287gb (~10.7m inodes) worth of data from a 630 to a clean volume on a 760. To schedule the downtime, I need to know if anyone has ballpark figures for calculating the tiem for a volcopy. The filers are well connected (gbit ether on both of them), and I can sacrafice big chunks of performance to make this go in a timely manner. (the 100mbit interfaces are inactive as well, if those should just be strapped together) Later on, we're looking at trying to snapmirror 536gb (~17.2m inodes) spread pretty evenly across two volumes between two locations. Are there numbers available on the approximate overhead of the snapmirror protocol (delta bytes * overhead), and how often it would be feasible to run it (I'd prefer something in the neighborhood of 5-10 minutes). The best numbers I could find on NOW were "36GB/hr will drive the cpu to 100%" without any reference to the filer model, cache sizes (the filer does do read-aheads, right?), inode density, etc... thanks. ..kg..

1 0

RE: "Wac"-ing the WAFL?
by Anders Ljungberg S (QDT) 30 Dec '99

30 Dec '99

We have 9 F760's in 3 clusters and 3 mirrors. They are all fully populated with shelfs ( 80 shelfs/~8TB in total ). I dont know the wack time because a never had to run a single one. The system was taken into production 9 months ago. So in our case it's not very common. Anders -----Original Message----- From: E Hunter [mailto:eric@gigaton.com] Sent: Thursday, December 30, 1999 4:28 AM To: toasters(a)mathworks.com Subject: Re: "Wac"-ing the WAFL? How long does it take to wack a fully loaded 760? By fully loaded, I mean somewhere in the 1TB range, which I think is the upper limit of the 760.

1 0

RE: "Wac"-ing the WAFL?
by Anders Ljungberg S (QDT) 30 Dec '99

30 Dec '99

This is my personal expirience from a heavy NFS Site. Which lead to replacing 2x7000/600 Auspex Netservers with a F740. To get realy good write performace lets say that you have to use atleast 7 drives/raidgroup.The performance raises up to lets say ~14. 7 drives without parity ( raid 4 or 5 ) is not verry funny considering the risk of a failing drive , restore times etc. So, in the Auspex case i had to go for mirrored stripes. ( Very bad price/performance and i finaly grew out of my cabinets ) On a filer i spend money on one "extra" drive instead of seven. The 740 is just cruising with the load which lead to the fall of the Netservers. Not to even mention the decreased need of administration. Anders Ljungberg As Jeff said, Netapp's RAID-4 implementation is not performance limiting in most cases. Most of the time a filer with RAID-4 will provide better performance than an equivalent product without RAID or with RAID-0. So I'm sure you will be quite pleased in your tests, particularly against a (*snicker*) Auspex. (Be sure to compare the price and number of filesystems in each system configuration as well.) I seem to recall some public talk waay back when about Netapp possibly offering RAID-0 since basically all you have to do is remove the parity drive. You could simulate this yourself by manually failing the parity drive and letting the filer operate in degraded mode with just the data disks. However, I'm not sure that the code is actually optimized for this case, so there might be some other normal RAID-4 checks that get in the way of realizing any performance gains. Bruce

1 0

RE: Some fun things not to try at home
by Priebe, Jason 28 Dec '99

28 Dec '99

> Are you implying that you'd rather use 6 drives per shelf > rather than 7? No -- I'm just assuming that the 7th disk (the hot spare) is drawing little or no power. Perhaps this is a bad assumption. > > Is startup power the power required to "spin up" the > > disks? In that case, isn't that power only delivered once at boot > > time? > > In addition to this, if I remember correctly, the filer spins up disks > sequentially, so there shouldn't be more than one disk drawing startup > power at any one time. That's good to note. Jason Priebe WRAL OnLine http://www.wral-tv.com/

2 1

RE: Some fun things not to try at home
by Priebe, Jason 27 Dec '99

27 Dec '99

Hmmm... that's interesting. I don't want to be foolhardy here, but I just checked the power requirements for the Seagate ST39173W (9GB Barracuda) drives and the ST15150W (4GB Barracuda) drives that are in the filer's shelf now. ST39173W ST15150W 12V startup 1.7 amps 2.18 12V typical 0.5 0.83 5V startup 1.2 amps 1.0 5V typical 0.75 0.75 It appears to me that, except for the 5V startup, the 9GB drives require _less_ power than the 4GB drives. Is the 5V startup the limiting factor? Can the power supply not deliver 1.2 amps at 5V to six drives? Is startup power the power required to "spin up" the disks? In that case, isn't that power only delivered once at boot time? Do the disks ever spin down in an active system? I really don't mean to be confrontational. We're just desperate to get some more life out of our F210 (we bought it new just two years ago). Jason Priebe WRAL OnLine http://www.wral-tv.com/ > -----Original Message----- > From: Bulfer, David [mailto:david.bulfer@netapp.com] > Sent: Wednesday, December 22, 1999 5:30 PM > To: 'Priebe, Jason'; toasters(a)mathworks.com > Subject: RE: Some fun things not to try at home > > > Please be careful. I do not believe that your StorageWorks > shelf has enough > power for 9G drives. Buying a new supply from Compaq would > not solve the > problem either. Your results could be quite unhappy. > > David Bulfer > Director of Platform Engineering > Network Appliance, Inc. > > -----Original Message----- > From: Priebe, Jason [mailto:priebe@wral-tv.com] > Sent: Wednesday, December 22, 1999 1:22 PM > To: toasters(a)mathworks.com > Subject: RE: Some fun things not to try at home > > > > -----Original Message----- > > From: Bruce Sterling Woodcock [mailto:sirbruce@ix.netcom.com] > > If you find out something > > useful in your filer experience, plesae pass it along, even > if it's of > > the "I didn't read the manual, so I did this and it > crashed" variety. > > While we're airing dirty laundry -- a couple of months ago, I > mentioned > to the list that we wanted to try to replace our F210's 4GB SCSI disks > with 9GB disks (full shelf, no more shelves available from NetApp). > > I replaced two of the 4GB disks with 9GB ones, and all was well (of > course, the filer only used 4 of the 9GB, but we figured we'd dump > to tape, rebuild the file system, then restore once all the drives > were in place). > > So I'm opening the 3rd disk carrier, removing the flex circuit from > the disk drive (has anyone else ever tried to do this? How the #$%! > do you do it?) by prying with a screwdriver, and the screwdriver goes > right through the circuit. D'oh! > > I did find that I was able to call Compaq and order a new > circuit (only > $96 for that tiny little piece!). Still waiting on delivery, though. > I'll let you know how it goes. Wish I'd known you could still order > these parts -- I could have built a couple of my own carriers and had > them ready to go into the filer instead of failing a drive, > disassembling > the carrier, reassembling the carrier, and reinstalling. > > Jason Priebe > WRAL OnLine > http://www.wral-tv.com/ > >

3 2

NFAS-1400
by Russell Murphy 27 Dec '99

27 Dec '99

I picked up a rack mount cabinet at the local electronic surplus and while looking it over to strip, I discovered it was a toaster. Before I cannibalize it, is there documentation available? Network Appliances' support pages appear to be customer only.

1 0

Spam alert
by Clawson, Simon 27 Dec '99

27 Dec '99

looks like this address has been hit with a spammer. Can you block this? Simon Clawson ---------------------------------------------------------------------------- ------------------------- > -----Original Message----- > From: mbhdkp5(a)hotmail.com > Sent: 26 December 1999 18:44 > To: toasters(a)mathworks.com > Subject: FREE SEX, Free forever, forever free! From: 219.130.141.242 > ---------------------------------------------------------------------------- --------------------

2 1

Jump to page:

2025

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