toasters

toasters@lists.teaparty.net

3 participants
13514 discussions

RE: performance problem on a 740
by Knight, Jason 12 Jan '01

12 Jan '01

Are you using SnapMirror on these F740's? -----Original Message----- From: Bertrand HUTIN [mailto:bertrand_hutin@amdahl.com] Sent: Friday, January 12, 2001 9:32 AM To: toasters(a)mathworks.com; vitar_dalloken(a)amdahl.com Subject: performance problem on a 740 On one of my customer's 740 filer, the average load is 72% and reached 100%, with OPS around 200 to 400. On another one with about the same OPS, the average load is 42%. Net and disk input/output is almost the same on both filers. what could be wrong?

1 0

RE: Can volumes span FCAL controllers?
by Mankovsky, Mike 12 Jan '01

12 Jan '01

My apologies for this response. It was intended for a different mail thread. Mike -----Original Message----- From: Mankovsky, Mike [mailto:mike.mankovsky@netapp.com] Sent: Friday, January 12, 2001 8:06 AM To: 'Chris Thompson'; scl(a)sasha.acc.virginia.edu Cc: sirbruce(a)ix.netcom.com; jason.santos(a)onsemi.com; mphelps(a)cfa.harvard.edu; toasters(a)mathworks.com Subject: RE: Can volumes span FCAL controllers? I don't see the reason to get Jess involved at this point. Mike -----Original Message----- From: Chris Thompson [mailto:cet1@cus.cam.ac.uk] Sent: Friday, January 12, 2001 7:45 AM To: scl(a)sasha.acc.virginia.edu Cc: sirbruce(a)ix.netcom.com; jason.santos(a)onsemi.com; mphelps(a)cfa.harvard.edu; toasters(a)mathworks.com Subject: Re: Can volumes span FCAL controllers? Steve Losen <scl(a)sasha.acc.virginia.edu> wrote: > [in response to Bruce Sterling Woodcock <sirbruce(a)ix.netcom.com>] > > > > > That's not entirely true -- you can span a volume across controllers, > > > just dont span RAID groups across controllers. > > > > When a volume has 2 RAID groups, is the NVRAM split among > > RAID groups? How are CPs done? > [ Good introduction to consistency points for newbies ] > > So in answer to your question, the NVRAM is shared by all volumes > and raid groups because it is a log of incoming write requests, > not a disk buffer cache. But to resume the original thread, the question is whether or not the writes done as part of CPs are clustered in a way that helps to reduce the overheads of switching between FCAL controllers. CPs for different volumes are logically distinct operations, but are in practice synchronised, either by the 10-second clock or by NVRAM filling up. In saying that one should avoid spreading a volume over multiple controllers, but can have different volumes on different controllers, the assumption is that writes (mostly) occur first to one volume, then to another. If the same is to apply to RAID groups, then the assumption is that the writes associated with taking a CP on a single volume are (mostly) clustered by RAID group. This sounds entirely reasonable, given that the filer tries to write whole stripes, or at least stripes in which as many planes as possible are being updated. Chris Thompson University of Cambridge Computing Service, Email: cet1(a)ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QG, Phone: +44 1223 334715 United Kingdom.

1 0

RE: Can volumes span FCAL controllers?
by Mankovsky, Mike 12 Jan '01

12 Jan '01

I don't see the reason to get Jess involved at this point. Mike -----Original Message----- From: Chris Thompson [mailto:cet1@cus.cam.ac.uk] Sent: Friday, January 12, 2001 7:45 AM To: scl(a)sasha.acc.virginia.edu Cc: sirbruce(a)ix.netcom.com; jason.santos(a)onsemi.com; mphelps(a)cfa.harvard.edu; toasters(a)mathworks.com Subject: Re: Can volumes span FCAL controllers? Steve Losen <scl(a)sasha.acc.virginia.edu> wrote: > [in response to Bruce Sterling Woodcock <sirbruce(a)ix.netcom.com>] > > > > > That's not entirely true -- you can span a volume across controllers, > > > just dont span RAID groups across controllers. > > > > When a volume has 2 RAID groups, is the NVRAM split among > > RAID groups? How are CPs done? > [ Good introduction to consistency points for newbies ] > > So in answer to your question, the NVRAM is shared by all volumes > and raid groups because it is a log of incoming write requests, > not a disk buffer cache. But to resume the original thread, the question is whether or not the writes done as part of CPs are clustered in a way that helps to reduce the overheads of switching between FCAL controllers. CPs for different volumes are logically distinct operations, but are in practice synchronised, either by the 10-second clock or by NVRAM filling up. In saying that one should avoid spreading a volume over multiple controllers, but can have different volumes on different controllers, the assumption is that writes (mostly) occur first to one volume, then to another. If the same is to apply to RAID groups, then the assumption is that the writes associated with taking a CP on a single volume are (mostly) clustered by RAID group. This sounds entirely reasonable, given that the filer tries to write whole stripes, or at least stripes in which as many planes as possible are being updated. Chris Thompson University of Cambridge Computing Service, Email: cet1(a)ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QG, Phone: +44 1223 334715 United Kingdom.

1 0

performance problem on a 740
by Bertrand HUTIN 12 Jan '01

12 Jan '01

On one of my customer's 740 filer, the average load is 72% and reached 100%, with OPS around 200 to 400. On another one with about the same OPS, the average load is 42%. Net and disk input/output is almost the same on both filers. what could be wrong?

1 0

RE: Can volumes span FCAL controllers?
by Knight, Jason 12 Jan '01

12 Jan '01

What happens in a volume that is spanned across two controllers when one controller fails? Does that appear as a double disk failure (assuming more than one disk is on the other controller)? Jason Knight

1 0

RE: performance problem on a 740
by Bond, Andrew 12 Jan '01

12 Jan '01

There may be nothing wrong, perhaps they just need a faster filer. If the applications are doing different things then the load on the filer will vary. You cannot use OPS as a guide unless you know exactly what the ops are. You should open a call with NetApp support to make sure though. Regards, Andrew Bond Network Appliance UK Consulting Systems Engineer Waterview House, 1 Roundwood Avenue, tel +44 (0)20-8756-6722 Stockley Park, mobile +44 (0)7801-383566 Middlesex, UB11 1EJ fax +44 (0)20-8756-6701 http//www.netapp.com Get answers NOW! - NetApp On the Web - http//now.netapp.com Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email. Please advise immediately if you or your employer does not consent to Internet email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of my company shall be understood as neither given nor endorsed by it. > -----Original Message----- > From: Bertrand HUTIN [mailto:bertrand_hutin@amdahl.com] > Sent: 12 January 2001 15:32 > To: toasters(a)mathworks.com; vitar_dalloken(a)amdahl.com > Subject: performance problem on a 740 > > > On one of my customer's 740 filer, the average load is 72% > and reached 100%, > with OPS around > 200 to 400. > On another one with about the same OPS, the average load is 42%. > > Net and disk input/output is almost the same on both filers. > > what could be wrong? > > >

1 0

Re: F740 + Cajun gigabit switches
by Jim VandeVegt 12 Jan '01

12 Jan '01

I'm looking for sites that have experience running Gb ethernet through Lucent Cajun switches, particularly the P550. In our environment, we have two F740 filers. One is the production machine that all the clients connect to (call it F1), the other (call it F2) is a snapmirror backup. When both F1 and F2 are connected to the public network with Gb links, we experience intermittent connection closings on large transfers between the filers. Network captures indicate reset (RS) frames being sent one way or the other closing the connection. We don't see this happen when F2 is on a 100 Mbps FDX connection. We also have not been able to reproduce the problem running snapmirror over a Gb private link between the filers. ===== Jim VandeVegt vandevegt(a)yahoo.com vandevegt(a)avaya.com http://members.xoom.com/vandevegt Toilet paper should come off the front! __________________________________________________ Do You Yahoo!? Yahoo! Photos - Share your holiday photos online! http://photos.yahoo.com/

1 0

RE: Net Stats, where from?
by rbonine＠millermartin.com 12 Jan '01

12 Jan '01

It's good to know we are not alone ;) --Roger "Muhlestein, Mark" <mark.muhlestein(a)netapp.com> on 01/11/2001 12:20:14 PM To: Roger Bonine/M&M@MMIS cc: Subject: RE: Net Stats, where from? Actually, there are quite a lot. And we have plans to add this feature for CIFS. I've updated the RFE we use to track this to add your request. Mark Muhlestein -- NetApp CIFS engineering -- mmm(a)netapp.com > -----Original Message----- > From: rbonine(a)millermartin.com [mailto:rbonine@millermartin.com] > Sent: Thursday, January 11, 2001 9:04 AM > To: toasters(a)mathworks.com > Subject: RE: Net Stats, where from? > > > > > This would be very handy for our site as well. (I get the > feeling that not many > people with NT-only sites such as ours buy filers.) > > --Roger > > > > > > "Todd C. Merrill" <tmerrill(a)mathworks.com> on 01/11/2001 09:36:44 AM > > Please respond to "Todd C. Merrill" <tmerrill(a)mathworks.com> > > To: toasters(a)mathworks.com > cc: (bcc: Roger Bonine/M&M) > Subject: RE: Net Stats, where from? > > > > On Thu, 11 Jan 2001, Witham, John wrote: > > > http://now.netapp.com/NOW/download/tools/ntaptop/ > > > > Display Top NFS Clients (netapp-top.pl) > > Note: This tool is unsupported - NetApp offers no support for it > > May I implore NetApp to make available similar capability to > that of `nfsstat -l` for CIFS clients? Maybe some hook into the pktt > capability to show per host packet/byte counts? These filers, after > all, are multi-protocol fileservers. Half the time for me it's one or > more Windows hosts that give me headaches in terms of filer load. I'd > love to be able to do ALL these functions: > > nfsstat -z > <wait> > nfsstat -l > > cifsstat -z > <wait> > cifsstat -l > > opsstat -z > <wait> > opsstat -l > > Until next time... > > The Mathworks, Inc. 508-647-7000 x7792 > 3 Apple Hill Drive, Natick, MA 01760-2098 508-647-7001 FAX > tmerrill(a)mathworks.com http://www.mathworks.com > --- > > > > > > >

2 1

RE: disk_fw_update issue with 840-5.3.7R2
by Thomas.L.White＠chase.com 11 Jan '01

11 Jan '01

Oliver: Are you saying that if I had specified the drive - I would not have had to failover ? I tried it two days ago with 6.0.1 on an 840 that belongs to a cluster. The initial insertion of the new drive caused it's own disk_fw_update to occur on the filer that the drive belonged to. The partner however complained at me the rest of the day. I tried to "update" without any precautionary steps and the partner asked me if the owner was at the ok prompt or not - so I spoke with Reena at NetApp tech support. As per her advise - I failed over and did the update then gave back. It seems excessive to me that the partner felt the need to fw_update a disk that was already updated. Seems that it would be a simpler task to make him realize the news. - granted this was my first encounter with the dreaded "Spare Drive" replacement causing such havoc - but I will do my bit in the future to be sure that NetApp knows the model and serial number of the drives I want to replace. tw "Krause, Oliver" <OKrause(a)netapp.com>@mathworks.com on 01/08/2001 03:50:31 PM Sent by: owner-toasters(a)mathworks.com To: "'Tom \"Mad Dog\" Yergeau'" <MadDog(a)fool.com>, "'toasters(a)mathworks.com'" <toasters(a)mathworks.com> cc: Subject: RE: disk_fw_update issue with 840-5.3.7R2 As i said you can upgrade a single disk with "disk_fw_update <disk>" without service disruption while this disk is a spare. Again, you have to specify the disk. Also does a disk firmware upgrade not take 2 minutes per disk if you use disk_fw_update. Oliver BTW: ONTAP 6.0 and higher is supposed to do a automatic update when you insert a disk with older firmware (have not tried it yet). Cool, isn't it? > -----Original Message----- > From: Tom "Mad Dog" Yergeau [mailto:MadDog@fool.com] > Sent: Montag, 8. Januar 2001 17:39 > To: 'Krause, Oliver'; 'toasters(a)mathworks.com' > Subject: RE: disk_fw_update issue with 840-5.3.7R2 > > > Just upgrading one disk still takes downtime for the whole > system. The > filer may be up, but if I can't access my data, effectively, > it's down. > > "Downloading disk firmware can be disruptive to the system. > It involves spinning down and then spinning up all > affected disks, and suspending all disk I/O to all disks while > that is happening. This may take up to 2 minutes to complete. > This delay may cause clients to hang, and may cause > CIFS clients to drop sessions." > > It should be noted that this notice can be misleading. It > actually takes up > to two minutes PER TYPE OF DISK to complete a disk_fw_update. > So if you > have two models of 18 GB drives and two models of 36 GB > drives, if can take > up to eight minutes to flash the firmware. Been there, done > that. Was a > little surprised the first time. :) > > Since replacement disks often ship with firmware different > than the level > currently being used on a filer, it would be nice if > (a) NetApp checked autosupport logs and pre-flashed the > firmware for you > before sending you the disk > (b) You could flash a single disk w/o disrupting all I/O for > the entire > filer. Especially if the disk is marked as a spare and > therefore not part > of any volume. > > If NetApp wanted to get *really* ambitious (I dunno, for ONTAP 6.5 or > something), this would be wild: flash firmware on a volume > one disk at a > time, and keep the volume online. While each disk is being > flashed, use the > parity disk to serve data, and keep track of writes somewhere > to get the > flashed disk back up to synch when the flash is done. > > It's too bad NetApp doesn't offer mirroring, since if you had > mirrored sets > this would be a real piece of cake. Just flash half a mirror > while the > other half served data. > > Online disk_fw_update would be a big help in five 9s environments. > > MD > > > > -----Original Message----- > From: Krause, Oliver [mailto:OKrause@netapp.com] > Sent: Saturday, January 06, 2001 5:56 AM > To: 'Premanshu Jain'; 'toasters(a)mathworks.com' > Subject: RE: disk_fw_update issue with 840-5.3.7R2 > > > You can specify which disk to upgrade. Try > > disk_fw_update <diskname> > > This works while the system is running. > > Oliver > > > -----Original Message----- > > From: Premanshu Jain [mailto:PrJain@shastanets.com] > > Sent: Freitag, 5. Januar 2001 23:31 > > To: 'toasters(a)mathworks.com' > > Subject: disk_fw_update issue with 840-5.3.7R2 > > > > > > I have a 840 cluster. Whenever I add new disks, the cluster > > starts throwing > > 'doenrev disk firmware' error messages. NetApps suggests doing a > > disk_fw_update on the disk and strictly says "NO need to Halt > > a filer'. > > Reality is totally different. The command specifically asks > > you wether the > > cluster partner is on Halt-OK prompt or not? > > > > The workaround I am doing, intitiate a takeover and do the > > disk_fw_update. > > However, this still needs to do a takeover/giveback and 2 minutes of > > downtime...Adding/replacing bad disks is a regular activity... > > > > Any suggestions..? > > >

1 0

72 GB rebuild times
by Tom "Mad Dog" Yergeau 11 Jan '01

11 Jan '01

Has anyone had a 72 GB drive fail yet? If so, can you let us know how long the rebuild took and how many drives are in the volume? I'm particularly interested in how long it takes to rebuild a 72 GB drive in a 6-7 disk and a 10-14 disk volume. Thanks, MD

1 0

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