toasters

toasters@lists.teaparty.net

2 participants
13529 discussions

Re: hard link .snapshot files
by Mark Muhlestein 05 Nov '98

05 Nov '98

As I'm sure you've already discovered, you can't create a hard link to a file in a snapshot. The reason is that snapshots are read-only, but it is necessary to increment the number of links in the linked-to inode. You can use a symbolic link, of course, but it would point to whatever was at that location. To prevent the snapshot from being recycled, you could do a "snap rename" of hourly.0 or whatever to something else. Mark -----Original Message----- From: Jim Harm <harm1(a)llnl.gov> To: toasters(a)mathworks.com <toasters(a)mathworks.com> Date: Thursday, November 05, 1998 8:56 AM Subject: hard link .snapshot files >We have the need to occasionally recover a LARGE file and we would like to >recover it quickly without making the massive increase in space usage that >would result if we merely copied the file from .snapshot to the old directory. > >Can we build a hard link from the old directory to the >.snapshot/hourly.#/Large.file and how will that be protected with the next >snapshot? > >I don't want to screw anything up. Our users ARE rocket scientists. > >}}}===============>> LLNL >James E. Harm (Jim); jharm(a)llnl.gov >(925) 422-4018 Page: 423-7705x57152 > > >

2 1

Re: can't see ~snapshot directory
by Tom Reingold 05 Nov '98

05 Nov '98

On Thu, 05 Nov 1998 08:01:57 CST, David Drum <david(a)more.net> wrote: > I find it disgusting that NetApp caters to Microsoft's ineptitude. Let's be fair. snapshots are strange. They are links to files, but they appear to be duplicates. We had to provide our unix users with a new "du" command so it wouldn't traverse snapshots. How could Microsoft have foreseen this problem and written their apps accordingly? -- Tom Reingold tommy(a)bell-labs.com http://www.bell-labs.com/~tommy The problem with this world is that there's too much apathy. But I don't really care.

6 7

Re: can't see ~snapshot directory
by Tom Reingold 05 Nov '98

05 Nov '98

On Thu, 05 Nov 1998 11:05:26 PST, "Chad H. Tudor" <chtudor(a)stanford.edu> wrote: > : find . -name .snapshot -prune -o -print > > While this is really helpful; the find would not be for me, as I do > understand how find operates and more importantly I understand the command > man 'splat'. What I do have is a shop full of folks who do not and it > would be easier if I could point them to a specific biniary. I betcha it would be trivial to take the source code for gnu find and put in a rule never to traverse a directory by a given name. -- Tom Reingold tommy(a)bell-labs.com http://www.bell-labs.com/~tommy Bell Labs, the Research and Development unit of Lucent Technologies Murray Hill, NJ, 07974-0636 US

1 0

Re: can't see ~snapshot directory
by Tom Reingold 05 Nov '98

05 Nov '98

On Thu, 05 Nov 1998 09:54:20 PST, "Chad H. Tudor" <chtudor(a)stanford.edu> wrote: > :Let's be fair. snapshots are strange. They are links to files, but > :they appear to be duplicates. We had to provide our unix users with > :a new "du" command so it wouldn't traverse snapshots. > > This raises another question. > > Would there be a find laying about that ignores same? find . -name .snapshot -prune -o -print

2 1

Re: can't see ~snapshot directory
by David Drum 05 Nov '98

05 Nov '98

Quoth Mark Muhlestein: > Huh? This change only affects CIFS clients. *ouch* I guess I missed that in the original poster's email. Sorry. Regards, David K. Drum david(a)more.net -- It's hard to be bored when you're as stupid as a line. [1] Reality has a tendency to be so uncomfortably real. [2] When you proceed deliberately, mistakes don't cascade, they instruct. [3] What you notice becomes your life. [4] [1] Vernor Vinge [2] Neil Peart [3] Stewart Brand [4] Michael Chitwood

1 0

Re: can't see ~snapshot directory
by Tom Reingold 05 Nov '98

05 Nov '98

On Thu, 05 Nov 1998 11:21:38 CST, David Drum <david(a)more.net> wrote: > I will admit that file system theory is not that different from UNIX > to PC. Therefore you are correct. My question then is: Why is this > only a problem now? How were UNIX users able to manage with one default > behavior all this time? Why can't the CIFS users just RTFM and configure > the NetApp so that it works the way they need? NetApp is thumbing its > nose at the installed base of (UNIX) users, who will now have to restore > the old default configuration. Just my Microsoft rant of the day. We mount each user's home directory as /home/whatever, so it does actually cause a problem (with things such as du). That is to say it's a problem for both PC and UNIX users. You could argue that the solution should be on the client side, but until that comes, it's nice that the server offers us an option. But you can't argue that Windows is more broken than UNIX in this case.

1 0

Re: Dual home fiber
by Gilbert Smith 05 Nov '98

05 Nov '98

At 10:13 AM 11/5/98 +0000, Algar, Christopher [PAI01:8881:EXCH] wrote: >Greetings. > >I'm having a bit of difficulty configuring a F760 with two fiber cards on >different sub nets. >It's been running fine with ifconfig f3 `hostname`-f3 netmask 255.255.240.0 >I've tried adding a similar line to etc/rc for f1 explicitly stating the IP >address instead of hostname, installing the additional card in slot 1 and >rebooting. Also tried installing the card with no changes to rc and >rebooting. >The result is the same, the system panics and fails to boot. > >Can anyone help? I would try floppy boot using option (no /etc/rc, I think #2 from the menu). Then run command, "sysconfig -v" and examine the output. Do you see both cards listed in the sysconfig output and are the two cards identical? Now run command "setup" and answer the questions. Setup will keep a backup copy of all your /etc/?? config files like (/etc/rc, /etc/exports, /etc/hosts, etc.). But setup should choose the appropriate mediatype and build the correct rc file for your hardware. There is a mediatype argument to the ifconfig command. The panic messge you receive might help. Hope this helps, Gil

1 0

Re: can't see ~snapshot directory
by sirbruce＠ix.netcom.com 05 Nov '98

05 Nov '98

On 11/05/98 11:39:35 you wrote: > > >On Thu, 05 Nov 1998 08:01:57 CST, > David Drum <david(a)more.net> wrote: > >> I find it disgusting that NetApp caters to Microsoft's ineptitude. > >Let's be fair. snapshots are strange. They are links to files, but >they appear to be duplicates. We had to provide our unix users with >a new "du" command so it wouldn't traverse snapshots. > >How could Microsoft have foreseen this problem and written their apps >accordingly? The real problem, IMHO, is that the snapshots should be treated as their own seperate partition, and be mounted as such, rather than inserting .snapshot directories. Of course, this does make it a little more difficult to explain to the user where to go to get their files, and with NFS it necessitates lots of new mounts. You could just turn the snaps off and mount it seperately and test this yourself; I think it will work. It also avoids the problem of copying the old file over it's newer copy with the same inode (in UNIX). Bruce

1 0

Nifty symlink trick
by Stephen C. Losen 05 Nov '98

05 Nov '98

This may be old news, but I discovered it while playing around with a F630 running 5.0.2. The netapp filer itself follows symlinks. I guess that shouldn't be a surprise, but, technically, I don't think that a filer has to follow symlinks. In NFS, the CLIENT follows the symlink, not the server. The server simply passes to the client the contents of the symlink. So one could design a filer that does not itself follow symlinks and it would still be a correct NFS server. (CIFS is another story.) Our F630 has two volumes: /vol/vol0 (the root volume) and /vol/vol1, which NFS clients mount on /h2. Just for fun, I created this symlink in /vol/vol0 /vol/vol0/h2 -> /vol/vol1 So now on the netapp, /h2 -> /vol/vol1 I discovered I can now use /h2 in config files wherever I would ordinarily use /vol/vol1. In /etc/exports I have successfully exported /vol/vol1 like this: /h2 access=foo.bar.EDU And this command works on NFS clients for mounting: mount home2.Virginia.EDU:/h2 /h2 And it also works in the quotas file. /h2/users/loginid user@/h2 100M 100K Note that this trick only works where /vol/vol1 would be appropriate. For example I still have to use "quota on vol1", but I can change the volume name from vol1 to h2, to fix that up. (and then change my symlink to /vol/vol0/h2 -> /vol/h2) An advantage to using this trick occurs to me. If you change a volume name, you may be forced to change quite a few config files. If you use a symlink in the config files, you can rename a volume and then just change the link. Also, pathnames on clients, such as /h2/foo/bar will also work on the server. So then you don't have to always translate paths in your head from "client-style" to "server-style" when running a command on the server, such as "quota report /h2/foo/bar" Disclaimer: I don't know if Network Appliance will guarantee this behavior in future releases. Steve Losen scl(a)virginia.edu phone: 804-924-0640 University of Virginia ITC Unix Support

2 1

RE: Changed DLT 4000 for DLT 7000 - how to get new storage densit y?
by Thomson, Michael (UK, GIS ISM Group for R&D) 05 Nov '98

05 Nov '98

Apologies to you all, but this was not a problem at all. I must have examined a tape that hadn't been written to. Last nights backups behaved as expected, writing at the full density capability of the drive. <sheepish grin> Mike -- \\|||// Minds are like parachutes; (@ @) They only work when open. *---oOO-(_)-OOo------------------------------------* * Michael Thomson Email: mt53590(a)ggr.co.uk * * Unix Specialist Internal: 782 2369 * * Glaxo Wellcome External: 01438 762 369 * * Stevenage, UK Fax: 01438 762 461 * * .oooO Oooo. Mobile: 0498 725 047 * *--( )---( )-----------------------------------* \ ( ) / (Quote: Sir James Dewar) \_) (_/ > ---------- > From: Thomson, Michael (UK, GIS ISM Group for R&D) > Sent: 05 November 1998 11:49 > To: 'support(a)netapp.com' > Cc: 'toasters(a)mathworks.com' > Subject: Changed DLT 4000 for DLT 7000 - how to get new storage > density? > > Hi there. > > Last weekend we upgraded our DLT4700 (single DLT4000, 7 slots) for a > Brece-Hill Q2.15 (two DLT 7000, 15 slots). > We're reusing the CompacTape IV cartridges, but the new drives seem to > recognise the old tapes as being written by a DLT4000, and don't use the > higher density that the DLT7000 is capable of. > > Is there an easy way to re-initialise the tapes that'll allow us to use > the maximum storage capacity? > > Details follow: > > First we check the status of the drive: > > Checking pre-dump status with: rsh -n <filer> mt -f nrst0a status > Tape drive: Digital DLT7000 > Status: ready, write enabled > Format: 85937 bpi 70 GB (w/comp) > fileno = 0 blockno = 0 resid = 0 > > This implies that it'll use the high density we're after... > > Then we do the dump > > rsh -n <filer> dump 0ufbB urst0a,urst0a,urst0a,nrst0a 63 30000000 / > > (Yes, I know the block count is low - we're trying to fix the problems > without causing any EOT fun & games) > > Now after the dump is done, we load the tape into the drive, which reports > > A: Idle C 20.00G > > This says to me that the tape is still recorded at the DLT4000 density, > albeit with compression. > > Any suggestions? > > Cheers, > Mike > -- > \\|||// Minds are like parachutes; > (@ @) They only work when open. > *---oOO-(_)-OOo------------------------------------* > * Michael Thomson Email: mt53590(a)ggr.co.uk * > * Unix Specialist Internal: 782 2369 * > * Glaxo Wellcome External: 01438 762 369 * > * Stevenage, UK Fax: 01438 762 461 * > * .oooO Oooo. Mobile: 0498 725 047 * > *--( )---( )-----------------------------------* > \ ( ) / (Quote: Sir James Dewar) > \_) (_/ > >

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