toasters July 1998

toasters@lists.teaparty.net

51 participants
51 discussions

Re: ndmpcopy?
by Darrell Root 23 Jul '98

23 Jul '98

Ok, so the behavior acts like this: 1) On client A, which had the filesystem mounted before the ndmpcopy, the new data does not show up in directory X. 2) On client B, which mounts after the ndmpcopy starts, the data shows up. 3) On client A, after you unmount and remount, the data shows up. I have a theory based on the nfs protocol which might explain the behavior. Have your flamers ready because this could be wrong wrong wrong. Keep in mind that I've never actually read any nfs client source code...so set your expectations accordingly... When you do a ls on a directory you've never accessed before from a client, you get the data block for the directory and put it in your buffer cache. Later, if somebody does an ls on the same directory on the nfs client, the client needs to see if the buffer cache's data is valid. It does this by executing a getattr() nfs call. If the timestamp on the server has not changed, then the client will think that it's buffer cache is valid and not actually get the directory data. (this is why latex takes so long in nfsv2 even if the whole thing is in your buffer cache, depending on how much latex stuff you have you might have to do 10,000 getattr() calls to validate your buffer caches-- whereas if you're running on local disk you _know_ the buffer caches are valid). So, if the ndmpcopy is not updating the timestamp on that directory until the end, the client will happily believe that the directory's contents (as listed incorrectly in the clients buffer cache) is correct. But if you unmount and remount, that invalidates that part of the buffer cache, causing the data to magically appear. And if you mount on another client after the ndmpcopy is in progress, you don't have an incorrect buffer cache entry. This theory explains the behavior. Here are experiments to increase confidence in the theory: 1) packet trace between the server and the client, to watch for the getattr(), but not the actual read of that directory. 2) Manually create a file in the directory during the ndmpcopy from a second client (which causes the directory timestamp to update, which will mean when the first client does it's buffer cache validation with getattr() it will realize that it's buffer cache is out-of-date. 3) Check the timestamp on the directory from client 1 before the ndmpcopy, and from both clients during the ndmpcopy. Of course, this isn't important enough for me to try these experiments. If this is true, then it's a simple netapp bug which can be fixed by updating the timestamp on the target directory at the beginning of the ndmpcopy. Of course, if it's wrong then it's wrong :-) Darrell Root rootd(a)nas.nasa.gov

3 2

INN 1.7.2+insync/Overview/5.0.1 F520-110
by Larry Rosenman-CyberRamp System Administration 23 Jul '98

23 Jul '98

Let me introduce myself. I am the Sr. System Administrator and all around chief tech-nerd for CyberRamp Internet Services in Dallas, TX. We are a ~16,000 user IS. We recently (~6-8 weeks ago) put up a F520-110 Filer running 5.0.1 software. When I try to run INN (InterNetNews) 1.7.2 and put it's overview files on the filer, I get Overchan hanging very regularly on a fcntl(F_SETLK,...)/fcntl(F_SETLKW,...) loop. My question for the assembled guru's is: Is anyone else seeing this? What other information should/could I supply to help? I can't afford to recreate this many more times, as recovery is a pain. Any help/advice would be appreciated! Larry Rosenman -- Larry Rosenman, Sr. System Administrator, CyberRamp Internet Services E-Mail: ler(a)cyberramp.net, http://www.cyberramp.net Voice: (214) 343-3333/(817) 461-8484 (Metro)/Fax: (214) 343-3727 U.S. Mail: 11350 Hillguard Rd, Dallas, TX 75243-8311

2 6

Disk usage script
by Madhu Chandra 23 Jul '98

23 Jul '98

Hi, Does anyone have a script that monitor the diskusage on netapp box and sends an email or sends a pager notification. We have: Solaris 5.5.1 and netapp F520 and F230 If you have one please email it to me Thanks Madhu Chandra (617)374-3418

5 4

subscribe
by Farouk Al-Shorafa 23 Jul '98

23 Jul '98

subscribe -farouk

1 0

Re: ndmpcopy?
by grant＠netapp.com 23 Jul '98

23 Jul '98

Hi, > > > I need to move about 56GB of data from one F520 to another. Does > > anyone have experience using ndmpcopy for that? It looks like it > > might fit the bill, but I'd be interested in hearing of any > > good-bad-indifferent experiences with it, and particularly what rate > > it can transfer data from one filer to another. > > Answer here will help me out as well... > > Gotta move 20GB from a little F210 to a new box... > > > I also need to move 36G from F220 to F520. > > Please post answer. > With Data OnTap 5.1 you now have options: If you are migrating a complete volume from one filer to another (or even copying one volume to another on the same filer) you can use SnapCopy(tm) which allows you to copy either the active filesystem or the active filesystem *plus* all the current snapshots. Its a licensed product and uses the vol copy command. NDMPcopy can copy complete volumes either locally or remotely, but it can also do sub-directories or quota trees and is useful if splitting a large volume into several smaller volumes. Although dump creates a snapshot to backup only the destination's active filesystem image is updated via NDMPcopy. I wouldn't count on seeing NDMPcopy achieve > 6 GB/hr (sustained) until sometime next year, when we *hope* to have some improvements. Cheers, Grant

1 0

Re: ndmpcopy?
by grant＠netapp.com 23 Jul '98

23 Jul '98

Hi Matt, > > Sorry, I should have been more clear. My filer contains a tree of data > that a web server serves up to the net. In a mass restore situation, I > want to put the web server up right away on that small amount of data, and > let the rest fill in in the background. > > The idea here was that once we kicked off the restore, some data would be > available right away, rendering the system 'up'. Proper balancing and > placement of files on the filesystem would even ensure that the most > important data came back first. > If that's your goal and you can identify what data is important to get back first it may be worth throwing this important data in a quota tree and backing it up separately. It may take more tape juggling to restore all your data, but you may be able to relax a little once you know you are serving the important data again. I say use a quota tree mainly to help with the backing up part, since in Data OnTap 5.1 we have optimized dumping full volumes and full quota trees. Cheers, Grant

1 0

RE: ndmpcopy?
by Kelly Wyatt 22 Jul '98

22 Jul '98

I recently moved 65gb from one filer to another using ndmpcopy. The first try was successful from an ndmpcopy standpoint, but I ran out of space on my target volume (duh!). However, succeeding tries were not successful - it crashed the *source* (production) filer. I see from a followup post that this is probably caused by the level of my dump... I also heard from netapp that this should be fixed in 5.01. I was running 4.3.1 on the source filer and 5.1 on the target filer. Due to unforeseen circumstances (let's just say it wasn't pretty), this filer is now at 5.01. I plan to try a copy again. By the way, we achieved about 6gb per hour on 100bT ethernet between 2 F630's. Kelly ------- Kelly Wyatt, saskww(a)wnt.sas.com <mailto:saskww@wnt.sas.com> Integrated Solutions Consulting SAS Institute Inc. / SAS Campus Drive / Cary, NC 27513 (919) 677-8000 X4279 (Voice) / (919) 677-4444 (FAX) http://www.sas.com <http://www.sas.com>

2 1

Re: ndmpcopy?
by fred.abounader＠vantis.com 21 Jul '98

21 Jul '98

I also need to move 36G from F220 to F520. Please post answer. ---}X-Pgp-Fingerprint: D8 24 CC E6 47 5F E4 60 BF B7 6E FA BF C7 6E C5 ---} ---}On Tue, Jul 21, 1998 at 01:37:58PM -0700, Jim Davis wrote: ---}> I need to move about 56GB of data from one F520 to another. Does ---}> anyone have experience using ndmpcopy for that? It looks like it ---}> might fit the bill, but I'd be interested in hearing of any ---}> good-bad-indifferent experiences with it, and particularly what rate ---}> it can transfer data from one filer to another. ---} ---}Answer here will help me out as well... ---} ---}Gotta move 20GB from a little F210 to a new box... ---} ---}-- ---}Mike Horwath Admin & Manager @ Vector Internet Services, Inc. ---}IRC: Drechsau HOME: drechsau(a)Geeks.ORG WORK: drechsau(a)yuck.net ---}Minnesota ISP with multiple T3s to the 'net? http://www.visi.com/ ---} Garbage In -- Gospel Out. - berkeley fortune(6) ---}

1 0

Re: No CPU stats with filer-mrtg (fwd)
by Tom Yates 20 Jul '98

20 Jul '98

thanks to all for the very kind offers to host the list. i'd sort of thought that this para i sent last week cleared that up, but apparently not, so: my replacement here at MathWorks has agreed to shoulder the SYSMAN's burden and continue with the list. again, though, thanks to the many kind offers of space for same - it's much appreciated. Tom Yates - Unix Chap - The Mathworks, Inc. - +1 (508) 647 7561 MAG#65061 DoD#0135 AMA#461546 1024/CFDFDE39 0C E7 46 60 BB 96 87 05 04 BD FB F8 BB 20 C1 8C ---------- Forwarded message ---------- Date: Fri, 17 Jul 1998 13:55:51 -0400 (EDT) From: Tom Yates <madlists(a)mathworks.com> To: Kendall Libby <fubar(a)mathworks.com> Cc: toasters(a)mathworks.com Subject: Re: No CPU stats with filer-mrtg [...] errrm, not quite. i hope that people are using http://teaparty.mathworks.com:1999/toasters . this will be valid indefinitely, afaik. [...]

1 0

Re: No CPU stats with filer-mrtg
by tjt＠netapp.com 19 Jul '98

19 Jul '98

> > I still have a problem with NFS ops. > > ... due to the fact that this filer unlike the others has been up > > 34 days and done 2,725,616,890 nfs ops I've fixed the filer-mrtg package so that it properly handles the NFS ops value when it rolls over. (I also fixed the perl path problem that was mentioned.) It will be updated on the NOW site shortly, but if you want a copy before that, just send me email. There is now a nifty "Subscribe" feature on the Tools&Utilities page of the NOW site, so if you want to be notified about new and updated tools, you should subscribe to it, and you'll get email whenever things are added or updated. ...Tim Thompson...Network Appliance, Inc...tjt(a)netapp.com...

1 0

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 July 1998