toasters

toasters@lists.teaparty.net

13515 discussions

rsync and nfs file systems
by Greg Kitch 24 Feb '00

24 Feb '00

We have experienced the following problem with the --delete option when using rsync to mirror data... When a directory on the source file system is replaced by a symbolic link with the same name, the directory is deleted from the destination file system by rsync by unlinking the directory without having recursively removed all files/directories under the parent (deleted) directory. The net effect on our local destination file system seems to be that this leaves unreferenced files and directories that must be cleaned up by fsck at some future date. We noticed this by accident simply because we decided to take an rsync-mirrored disk offline and fsck it to check its integrity. Our question is...What will happen to a wafl file system if rsync deletes a directory in this fashion? Does NFS deal with this properly or is it a potential issue? Thanks in advance. Greg Kitch UCLA SEASnet Computing Facility

6 7

RE: Disk drive technology
by Mohler, Jeff 24 Feb '00

24 Feb '00

I dont have any solid data on data/stripe/drivesize performance handy, but it is a wise move to not span raid-groups over multiple FCAL controllers. There will be write degredation involved with that that I have seen. With two F760s rebuilding a manually failed drive while doing nothing else, the machine without a stripe split across FCAL cards completed about 25% before the machine that did span FCAL cards for a raid stripe. mkfile tests from Solaris confirmed what Id heard..and saw in the reconstruct as part of my test. :-----Original Message----- :From: mds(a)gbnet.net [mailto:mds@gbnet.net] :Sent: Wednesday, February 23, 2000 5:04 PM :To: toasters(a)mathworks.com :Subject: Re: Disk drive technology : : :On Wed 23 Feb, 2000, "Bruce Sterling Woodcock" :<sirbruce(a)ix.netcom.com> wrote: : :> The article doesn't say, but bigger, faster disk drives :often draw more :> power. Thus, NetApp needs to have shelves that can support them :> first, before they can test them for reliability. There are :a lot of issues :> involving mechanical vibration and so on that have to be :assessed before :> such a drive could be qualified. It may be that this :particular drive is :> not appropriate for certain shelves. : :Well, yeah, so much so obvious. So how soon was that? 8) : :> More spindles is better performance, if you are read-heavy. :If your filer :> seems slow and it's reading constantly but the CPU is still :under 100%, :> more spindles could help. : :So much for conventional wisdom. However, my question was trying to :elicit something a little more, well, complete as a model of :performance. : :I mean FCAL is a loop technology, and filers can come with 1, 2 or 3 :loops. How much does performance differ when using 1 chain as opposed :to 3 balanced chains, using which drives, using what RAID set sizes and :volume mappings? How does performance get affected by the size of the :disks you use? : :See, if none of these things makes any difference to performance I'd :love to know why. If they *do* then I'd love to know how to build :faster filers. : :I'd be happy with an analytical model, or a set of tables taken from :either real lab testing with real filers, or simulator results. As it :stands though, all we have are rules of thumb that sound plausible. : :The same goes for backup performance. There's been so much discussion :on the list about both topics, but I can't honestly hand-on-heart tell :someone how best to set-up their filers. I'd write something if I knew :the answers myself. : :The best I can find is things like: :http://now.netapp.com/NOW/knowledge/contents/FAQ/FAQ_560.shtml : :I've been meaning to brush up on queueing theory so maybe I *will* try :to work something up myself on this. I'd just hate to reinvent :the wheel :if someone's done it already. I also don't know anything about the :buffers in the filer, on FCAL interfaces, the network interfaces, the :precise use of the WAFL, etc.. So my model might be under-informed. : :Here's a thought to mull over: if you have 20 18 gig drives on 2 loops :and 16MB nvram, will you see better or worse performance than :with 40 9GB :drives on 1 loop with 16MB nvram presuming your clients are all writing :as fast as they are allowed? Please disregard any "such configs aren't :shipped" objections as this is an in-principle hypothetical. Please do :consider the effect of RAID sets, and incremental disk-additions over :the lifetime of the filer in question. I picked writing as the task :because I wanted to point out also that there's an impact on the DOT :algorithms, based on the spindles, the RAID set sizes, and maybe the :chains too. : :I don't know the answer, but I'd like to. Without such models we all :have to purchase filers somewhat in the dark, and either pray that :we got it right, or expend more time than we could in evaluating our :purchases for suitability. : :One question - is anyone else interested in this or am I just shooting :my mouth off needlessly? : :>-- End of excerpt from "Bruce Sterling Woodcock" : :-- :-Mark ... an Englishman in London ... :

1 0

Disk drive technology
by mds＠gbnet.net 23 Feb '00

23 Feb '00

Having just seen: http://www.eetimes.com/story/OEG20000223S0007 I'm curious to know if they'll be appearing in NetApp's soon. I'm also curious to know if anyone has any more information to share regarding the effects on filer performance of the number of spindles, the number of chains, the size of the disks, etc, etc, etc. Everyone went a little quiet last time this came up. -- -Mark ... an Englishman in London ...

2 1

Re: Backing up a filer
by mds＠gbnet.net 23 Feb '00

23 Feb '00

On Mon 14 Feb, 2000, Lisa Paton <lcp(a)head-cfa.harvard.edu> wrote: Thanks for the review Lisa, word from teh front-line is far more valuable than marketing bumf. <snip> > supported. I liked the fact that both QR and Veritas used an ISO standard > tar format for the backup. This ruled Legato out early, since they use Hmmm, I suspect from what I've learned from Veritas and others, that you don't see tar format on the tapes used for the basic backups. The reasoning being that the backup software usually interleaves several streams of data from clients to ensure streaming onto tapes (to extend tape lifetimes as well as reduce the overall backup window). Cloning the tapes in some manner allows the backup software to reduce the chance of data loss AND also to rationalise the interleaving so you do get tape files with coherent tar format inside.. but some folks don't clone tapes, so they may well never get tar-readable tapes. I'd be happy to hear otherwise, but I wouldn't place too much value in that ISO standard tar formatting. I have to say that when I first heard, from Veritas Reps trying to sell their product to me, that the guys at CDC had nobbled gnu tar into an enterprise-class backup product (GPL notwithstanding) I just chuckled. Valid approach, but when you're backing up 100's of GB's or more, and spending through the nose for an enterprise-class backup product that manages every detail for you, the value of personally being able to grub through a tape using tar begins to pale a bit. > a proprietary format. I was impressed that Jim Ward (President of > Workstation Solutions) came out to meet with me personally, AND could > answer technical questions that the Veritas SE couldn't answer. WS impressed > me as a small, technically savvy company that was very responsive to me > during the evaluation period. I had a couple small glitches that were Now that sounds like NetApp - who still send out plenipotentiaries to find out what their customers are really doing, and really need. > Lisa Paton >-- End of excerpt from Lisa Paton -- -Mark ... an Englishman in London ...

4 4

How to Create ClearCase VOB on filer F720
by Amy Cheng (r46990) 23 Feb '00

23 Feb '00

Dear All, We've just purchased ClearCase and I want to create VOB on filer F720. But I met difficulties . When I run making vob command, there are some error showed. #cleartool mkvob -tag /proj/view -public /proj/test.vbs cleartool: Error: Unable to contact albd_server on host 'cptlfiler' >cleartool: Error: Unable to create view "/data/project/view/test1.vws". We have mounted a diretory of cptlfiler on /proj. Someone tell me that technical reason is that we can not do NFS mount over NFS. There is no way to created a view on filer without ClearCase. Have anyone met this qestion like me? How do you deal with it? if it not feasible to create vob on filer, we can't benefit from snapshot any more for vob. Sincere thank for you help. BR Amy

3 2

Re: level 2 oplocks and MS outlook
by Matt Day 22 Feb '00

22 Feb '00

Graham Knight wrote: > I think i am running into bug 20270: > > "Outlook users get access is denied (or similar) errors" > > The bug report states the cifs.lvl2_oplocks_cap should > be turned off in order to fix the problem. I have > about to 8000 CIFS users and only a handful have had > this kind of problem. What are the ramifications to > my other 7990 users if I turn off level 2 oplocks > to fix the 10 users that are having the problem? > > Should i just ship the 10 problem users to an NT server? Hi Graham, The ramifications of disabling level-2 oplocks depend on what your clients are doing. In a nutshell: if your clients share files frequently with each other then you should expect to see an increase in file reads when you disable level-2 oplocks because level-2 oplocks tell clients that it's OK to cache reads locally. Level-2 oplocks only come into play when there are multiple file openers, so disabling level-2 won't do much if you don't share files much. For example, if you had multiple NT clients doing compiles on a shared source tree, level-2 oplocks would allow the clients to cache the source files, speeding up subsequent compiles. Outlook is affected by level-2 oplocks because it sometimes opens the same file more than once. Level-2 oplocks aren't supported by Win9x, so those clients won't be affected at all. Here's Microsoft's KnowledgeBase article on NT oplocks: http://support.microsoft.com/support/kb/articles/Q129/2/02.asp Bug 20270 is fixed in 5.3.4R3P1 (available now) and 5.3.5 (coming soon). Matt

1 0

9GB disks
by Jason Lavoie 22 Feb '00

22 Feb '00

About a year ago, we bought a filer (720) with a single FCAL shelf, populated with five 9gig drives. This was about all the storage we needed at the time (for mail spool storage), but knew we could just add a couple of extra drives if we needed more storage. Now, we are looking at adding another large customer (another ISP) to our mail system, and needed a bit more storage. Adding another 9gigs or so would be fine. So, I figured we'd just fill the shelf with drives, and be happy. Netapp tells me that I can't buy 9GB drives anymore. They've proposed upgrading to a new shelf full of 18GB drives, and trading in the existing shelf. Needless to say, the price is exorbitant compared to just adding a couple 9GB drives as we had expected. I'm looking for alternatives. Is anyone selling old 9GB drives for Netapp? If so, please contact me directly so we can work something out. Does anyone have any information on how/where to source non-Netapp drives that will work in our system? Can I use 18GB drives in this shelf (and only use the first 9GB of each)? I've gotten conflicting answers about this from Netapp. Any other ideas? TIA -j -- "What is the sound of Perl? Is it not the sound of a | Jason Lavoie wall that people have stopped banging their heads against?" | jason(a)mint.net --Larry Wall in <1992Aug26.184221.29627(a)netlabs.com> |

3 2

Clustering and the 1.4 TB limit
by Derek Kelly 22 Feb '00

22 Feb '00

I'm sure this question has been posted in the past, but since I've only recently joined this list, I'm going to ask for my own greedy benefit... It is my understanding that to perform cluster failover between two NetApp boxes, the sum total of served data between the two boxes must not exceed a 1.4 TB. Since the two F760's we wish to cluster both have 1 TB each on them now, I can only conclude that we will not be able to take advantage of clustering. I have heard that NetApp plans to change this in the not-so-distant-future. If anyone has any "insider" <grin> info on this situation (i.e., an ETA date) it would be very much appreciated since we have a fairly immediate need to get these boxes clustered. Please feel free to correct any of the assumptions I'm operating under if i'm wrong. Best Regards, Derek Kelly Computer Systems & Support Group Genome Therapeutics Corporation www.genomecorp.com

1 0

RE: question about databases on NetApps
by Mohler, Jeff 22 Feb '00

22 Feb '00

Dont forget that you get to see your family again..no more annoying tuning of filesystems for performance. -----Original Message----- From: Bruce Sterling Woodcock [mailto:sirbruce@ix.netcom.com] Sent: Tuesday, February 22, 2000 12:01 PM To: toasters(a)mathworks.com; Mike Kazar Subject: Re: question about databases on NetApps ----- Original Message ----- From: Mike Kazar <kazar(a)icubed.com> To: <toasters(a)mathworks.com> Sent: Tuesday, February 22, 2000 11:22 AM Subject: question about databases on NetApps > Or are you just running one database host, and using the NetApp to manage > the database host's disk space. This is what most people do (I think). > If so, what are the advantages you get > from using the NetApp for disk space management instead of just using a > local disk? Well, the disk space is centrally managed, so you get your backup all in one place. You also get RAID for reliability, and snapshots, and often you get even better performance than you do with local disks (despite theory which might suggest otherwise). And your database host is easy to redesignate as needed; you don't have a directly attached disk array that you need to migrate with it. And of course, unique to the Netapp, adding additional disk space as needed is a quick and painless. Bruce

1 0

level 2 oplocks and MS outlook
by Graham C. Knight 22 Feb '00

22 Feb '00

I think i am running into bug 20270: "Outlook users get access is denied (or similar) errors" The bug report states the cifs.lvl2_oplocks_cap should be turned off in order to fix the problem. I have about to 8000 CIFS users and only a handful have had this kind of problem. What are the ramifications to my other 7990 users if I turn off level 2 oplocks to fix the 10 users that are having the problem? Should i just ship the 10 problem users to an NT server? Looking for answers from those in the know. :-) Thanks, Graham

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