toasters

toasters@lists.teaparty.net

13515 discussions

flock problems
by matthew zeier 11 Jun '98

11 Jun '98

This problem seems to happen about once a month. Wondering if it's a configuration probably on my end or just an issue with my filer. Mail is delivered to ~/.mailbox onto a F540 running 4.3.1D7. and clients are all Solaris 2.5.1. After about a month I start getting lots of calls from users (mostly elm users, some mutt users) about being unable to read their mailbox. A truss on the process shows it sleeping waiting for a flock on .mailbox. mutt uses dotlocking and seems to just hang doing an open(). Only solution seems to be a reboot. Any clues or pointers on where I should be looking? Thanks. -- matthew zeier -- mrz(a)3com.com -- 3Com EWD Engineering -- 408/764-8420 ...................................................................... "Y el mundo se mueve, mas rapido y mejor." - Fey

2 1

cvtusers question
by agy 11 Jun '98

11 Jun '98

>Is there a way to force CVTUSER to create the user list export both >alphabetically AND with ascending UID's? This way, as we add users to >the filer, we don't have to check for "free UIDs"? Or if we're not even >USING NFS (which we are not) and all we're using is the "cifs.home_dir" >option and nothing else, couldn't eveyone have the SAME UID??? There is a new version of cvtusers scheduled for July which will do what you want, as well as giving you the ability to edit the stuff before it's written to file. Check the NOW site toward mid-July in the Tools and Utilities section. The name of the new version is SecureShare Account Migrator. It is perfectly legal to give all users the same UID. Doing this results in FAT-style security, i.e. none. If that's what you want, go for it. Actually, since you're an NT shop, let me ask you, do you use a Domain Controller? If you do, the filer will be happy to use that for authentication as well, eliminating the conversion problem altogether. Alan -- --------------------------------------------------------------- Alan G. Yoder agy(a)netapp.com Network Appliance, Inc. 2770 San Tomas Expressway voice 408-367-3031 Santa Clara, CA 95051 fax 408-367-3451 ---------------------------------------------------------------

1 0

CVTUSER Questions...
by Weeks, Thomas 11 Jun '98

11 Jun '98

Hey all.. I was planning on using the CVTUSER utils to dump all my NT users into my FS230's (OT V4.3.4) /etc/passwd file in preparation for using the filer as a pure CIFS based home drive for our NT login users... (no NFS) My Q is... Is there a way to force CVTUSER to create the user list export both alphabetically AND with ascending UID's? This way, as we add users to the filer, we don't have to check for "free UIDs"? Or if we're not even USING NFS (which we are not) and all we're using is the "cifs.home_dir" option and nothing else, couldn't eveyone have the SAME UID??? Anyone been here before? Thanx... Tom

1 0

Re: filer-mrtg question
by Mark D Simmons 11 Jun '98

11 Jun '98

Well, I've got our local install of mrtg running with the non-cifs bits of filer-mrtg (I'm using an SGI for this monitoring task and already had an mrtg setup in place...so filer-mrtg was bit narrow-focus for my taste). Basically I lifted the first three parts of the filer-mrtg config file and run mrtg out of cron with them, for each of the filers I look after. I'm looking at the whole MIB for other useful things but my understanding of SNMP is still maturing. mrtg doesn't seem to be the right tool for some of the things I want to do though, along with the fact that MIB doesn't quite seem as simple as it could be. I know people will want me to justify that last statement, so here goes: There are raw figures for the number of, say, NFSPROC_READ calls since the stats were zeroed, AND there are percentages for the number of NFSPROC_READ calls against total NFS ops since stts began. This seems to be unuseful as it can be worked out from the raw figures, and SNMP was always meant to provide the raw data with the least processing at the agent side... What it practically means to me is that the MIB is large and I'm not very good at reading it. SNMP's umpteen different representations and translations are also causing me grief. If anyone's done more than rudimentary things with SNMP and mrtg, or anyother free snmp stuff, and feels like trumpeting their genius I promise to be suitably impressed when you show us how to do it properly! 8) Last time I counted there are 533 basic data (datums would convey the meaning better, but isn't a real word ;) in the MIB and some are only interesting when you perform some arithmetic on them and graph the results, and mrtg is confusing me as to how to get there from here. Probably my fault. Hey ho, back to the grindstone.

1 0

filer-mrtg question
by Alexei Rodriguez 09 Jun '98

09 Jun '98

Has anybody put this in place and have the nfsops code working? I see it is not there, yet graphs get generated for it. I just don't want to spin my wheels if others have it working already. Thanks. Alex

1 0

System RAM for F330 (or used NetApp stuff)
by Brian Marcotte 09 Jun '98

09 Jun '98

Lets try this one more time... NetApp tells me I need the following RAM for my F330: Buffered CAS DRAM, SIMM, 72 Pin, 8Mx36 60NS fast page mode, 4CAS 2K refresh, gold contacts This is not quite PC RAM, so I've not managed to locate a source. Any help locating this RAM would be greatly appreciated! Also, if anyone knows a source for used NetApp equipment, that would be a great help also. Specifically, I'd like parts suitable for a F330, or entire lower-end filers (F330, F210 ,F220, etc). I know that NetApp occasionally has refurbs, but I'd like more sources. Thanks! Note: I already know the dangers of using third party equipment in my filer. -- - Brian

1 0

subscribe toasters
by Madhu Chandra 09 Jun '98

09 Jun '98

subscripe toasters: can you subscribe me. Also can you tell me how to access and use it Madhu Chandra Sys Admin. Numeric Investors

1 0

5.0.1 or later?
by Daniel Quinlan 09 Jun '98

09 Jun '98

Anyone out there using 5.0.1 (or later, especially 5.0.1D1) with NDMP to directly attached SCSI tape drives? If you are, I'd like to hear whether or not you've had problems, or if any problems from earlier releases were resolved for you. Thanks.a - Dan

1 0

RE: Quota's, Qtrees, Old_WAFL, NewWAFL, etc...
by Weeks, Thomas 08 Jun '98

08 Jun '98

Yes... this sounds nice, but we do not use the filers tape drives (at this time) for backups... we are using NFS based Networker (Sun based Legatto product). I imagine after we get away from NFS and go to pure CIFS, we'll be moving our backups to full NT based (to keep any/all CIFS ACLs' intact with out files). Also, I imagine that your implimentation of dump does not allow for single file (or single directory) restorations does it? I realize that there are a couple products out there to that handle ACLS/UFS permissions backup & restoration via NPDM, but we are already running three backup variants right now and adding yet another is just not feasible. Thanx anyway... Thomas Weeks GTE SYSTEMS ENGINEER Testing-Integration Center "The Pit" Brooks AFB, DSN 240-5444 > -----Original Message----- > From: grant(a)netapp.com [SMTP:grant@netapp.com] > Sent: Friday, June 05, 1998 7:06 PM > To: toasters(a)mathworks.com > Subject: Re: Quota's, Qtrees, Old_WAFL, NewWAFL, etc... > > > Hi Thomas, > > Since you mention quota trees, I'd like to jump in and > discuss the benefits of quota trees within 5.1 (shipping > soon) and backup: > > My recommendation is that you should arrange your data in > quota trees whenever possible when running 5.1 for these > reasons: > > 1. You can have different security models running on a per > quota tree basis (ntfs, mixed (multi-protocol) or unix). > > 2. Pass I & II of dump has been optimized for full quota > tree (as well as full volume) backups. > > 3. dump has a new Q option that allows all non-quota tree > data to be backed up. This allows you to frequently > backup quota trees, but perhaps backup /etc and other > miscellaneous data on an infrequent basis and you don't > need to backup the whole volume to do so. > > Cheers, > Grant

2 1

Re: Quota's, Qtrees, Old_WAFL, NewWAFL, etc...
by grant＠netapp.com 06 Jun '98

06 Jun '98

Hi Thomas, Since you mention quota trees, I'd like to jump in and discuss the benefits of quota trees within 5.1 (shipping soon) and backup: My recommendation is that you should arrange your data in quota trees whenever possible when running 5.1 for these reasons: 1. You can have different security models running on a per quota tree basis (ntfs, mixed (multi-protocol) or unix). 2. Pass I & II of dump has been optimized for full quota tree (as well as full volume) backups. 3. dump has a new Q option that allows all non-quota tree data to be backed up. This allows you to frequently backup quota trees, but perhaps backup /etc and other miscellaneous data on an infrequent basis and you don't need to backup the whole volume to do so. Cheers, Grant

1 0

Re: Quota's, Qtrees, Old_WAFL, NewWAFL, etc...
by Keith Brown 05 Jun '98

05 Jun '98

> >I see that you can only use qtree's in the root... >??? - What is the real difference between quotas and qtrees? Filer qtrees are the entities on which you can slap tree quota restrictions in the /etc/quotas file. There are three types of quotas supported; User, group and tree. The former two are largely independent of qtree's, although they can be used in conjunction with them. The three can combine in relatively complicated ways and I'd recommend reading chapter 10 of the Systems Administration Guide for a full explanation. >??? - with 4.3.4 I can do non-root quota's for pure CIFS users right? If you mean quotas on individual Windows/CIFS users that are not associated with qtrees, then yes. Effectively, you would just define a regular user quota for each client. You could use group quotas and default quotas too (default quotas are just special cases of user quotas). >example: > >/cifs_home/<NT_username1> user 100M 20k >/cifs_home/<NT_username2> user 100M 20k >/cifs_home/<NT_username3> user 50M 10k Right. This would be fine. Just remember that the first field, the quota target field for a user quota is ultimately there to specify the UID of the user you are putting a quota on. If you are going to use a directory pathname in this field, the UID for the user quota comes from the owner of that directory. For this to work, the directory /cifs_home/<NT_username1> would have to have an ownership of whatever UID <NT_username1> was equated to via <NT_username1>'s entry in /etc/passwd (or NIS). >??? - These are just quotas on user directories right? No. User quotas are *not* specifically associated with the users home directory. User quotas expressed this way limit the users consumption of file and space resources on the root volume by default. If you want to limit the users space on another volume, you'd use the: /cifs_home/<NT_username1> user@/vol/foo 100M 20K syntax. >??? - I don't need quota trees for this type of individual CIFS user >quotas do I? Not if the above description is what you were looking for. > >The manual says that the user must have an entry in /etc/passwd. >??? - Is this true in a pure NT/CIFS file space? Yes. >??? - If so, Why? (no UID/GID?!) Because the user and group quota mechanism needs the files and directories created on the file systems to be owned by the UID's and GID's it's trying to impose its limits on for things to work. >??? - Will this change with 5.x? Not initially in 5.1 no. Sorry... >Lastly... >??? - Should I wait to impliment quotas until AFTER I go with 5.x or >will my 4.3.4 quotas and qtrees transition over ok to 5.x? Whatever system you come up with in 4.3.4 will transition just fine to 5.X as long as you keep doing things largely the same way. If you start creating multiple volumes in 5.X, then obviously some quota configurations may be effected by that. If you stop routinely mapping Windows/CIFS clients to valid UID's and GID's through /etc/passwd (which becomes perfectly feasible if you are using the NTFS compatible file system security in 5.1), then you will lose user and group quota-ing capability for those users, at least until we sort out some quota enhancements that will let you place quotas on users who are entirely defined by a Windows NT SID. However, tree quotas should keep on truckin' quite nicely, limiting even these kinds of users (the unmapped kind). Keith

1 0

Quota's, Qtrees, Old_WAFL, NewWAFL, etc...
by Weeks, Thomas 05 Jun '98

05 Jun '98

Hi all! Just some basic Q's... We're at OnTap 4.3.4 right now using the old WALF file system... I have implimented quota's yet and was wondering: I am planning on going to 5.x here soon (as soon as I can see how you all fair with it) along with it's full ACL's and implimentation of WALF... I am not currently using qtree's or quotas... I see that you can only use qtree's in the root... ??? - What is the real difference between quotas and qtrees? ??? - with 4.3.4 I can do non-root quota's for pure CIFS users right? example: /cifs_home/<NT_username1> user 100M 20k /cifs_home/<NT_username2> user 100M 20k /cifs_home/<NT_username3> user 50M 10k <etc...> (using PDC/BDC NT authentication for usernames) ??? - These are just quotas on user directories right? ??? - I don't need quota trees for this type of individual CIFS user quotas do I? The manual says that the user must have an entry in /etc/passwd. ??? - Is this true in a pure NT/CIFS file space? ??? - If so, Why? (no UID/GID?!) ??? - Will this change with 5.x? Lastly... ??? - Should I wait to impliment quotas until AFTER I go with 5.x or will my 4.3.4 quotas and qtrees transition over ok to 5.x? Please reply ASAP.. Thanx.. Thomas Weeks GTE SYSTEMS ENGINEER Testing-Integration Center "The Pit" Brooks AFB, DSN 240-5444

1 0

Backup strategy suggestions
by vance＠corp.theglobe.com 05 Jun '98

05 Jun '98

We have recently acquired an F230 which, once it is in production, host about 20GB of data. At the rate which new data is being added, we expect this to hit 80GB within 6 months. At that point, we expect to divide the load between a couple filers to maximize throughput. We will have a daily 8 hour window in which to perform backups. The average file size is about 20kB, and files are never modified or renamed, only created or deleted. Given this, does anyone have any words of wisdom about how to tackle backups? Thanks in advance for any advice you might have- Vance ......................... ........... vance huntley .. director of technology ............ theglobe.com . vance(a)corp.theglobe.com .........................

1 0

Soft Quotas
by Arvind K Aggarwal 05 Jun '98

05 Jun '98

Hi, Is there any way to set soft quotas? What I mean by soft quotas: If I give a user a quota of 100MB and he/she exceeds it, I don't want the system to die on him. Based on how much the excess disk usage is, I would a) Like to send him a message b) Give him/her a warning and then c) Disallow further disk usage. This is some what like the "GOD" program on the old VAX systems, if someone remebers those things :-); I have heard that it was pretty good. Thanks in advance Arvind Aggarwal

1 0

subscribe
by Elena Samsonova 05 Jun '98

05 Jun '98

1 0

Replication
by vance＠corp.theglobe.com 29 May '98

29 May '98

We would like to use some form of replication to mirror our NetApp 230 to another NetApp in a remote location. Has anyone on the list tried to do this before? Any suggestions on how to tackle this problem? Thanks in advance for any help you can provide- Vance BTW - If it matters at all, we are looking to replicate ~2M files totaling about 50GB. The WAN link would be fairly fast (> 45Mb). ......................... ........... Vance Huntley .. Director of Technology ............ TheGlobe.com . vance(a)corp.theglobe.com . http://www.theglobe.com .........................

3 4

subscribe
by David B. Ritch 28 May '98

28 May '98

subscribe toasters

1 0

User Quota Checking
by Mike Mueller 25 May '98

25 May '98

We've just turned on user quotas and I am trying to figure out a way to allow users to check their current disk usage and quota limit. Since the filer is running rquotad, this would seem to be the way to go. But alas, the filer rquotad apparently wants quota [path] instead of quota [uid], just like its command line quota report. It returns "NO QUOTA" when queried via a client's quota command. Has anyone figured out a way to allow a general user to check their disk usage and quota limits? It looks like there may be a way through the MIB, but how could this be easily made into a command line script that a user could run? Better yet, is there a way to allow the user to interact with the rquotad deamon that is already running on the filer to get their quota report? If not, why is it running? Thanks for your insight! -- Mike

3 3

command reference guide discontinued in 5.x
by Tom Yates 21 May '98

21 May '98

here's a little someting to take our minds off how nice NTP support would be... would anyone from netapp care to comment on the suggestion that the command reference guide has been discontinued in 5.x? would anyone else care to comment on whether they use the command reference guide? perhaps some might care to say that it's about twice as useful as the system administrators' guide, despite weighing half as much? i certainly think so. i'd even pay for it. i'd think it was a bit rich, given the price of the toaster, but i'd shell out folding money for it. it's just too useful. any chance we could have it back? 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

2 1

Re: don't run an F630 with the case open
by Mark D Simmons 21 May '98

21 May '98

Credit to our SE's for pointing this one out to us when they installed our F630's. I can see why this would drive you up the wall trying to diagnose a fail-to-boot when no-one's pointed out the change.

2 1

Re: don't run an F630 with the case open
by Guy Lauterbach 21 May '98

21 May '98

At 08:28 AM 5/21/98 -0700, Christina Sanfilippo wrote: >Ouch! That's got to hurt. I am sure Rich was glad he could fix your problem. > >For the F200/F300 and F500 series filers these filers can run with the case >open. >As you found out the F630 can't. Although some NetApp filers *can* be operated with the case open, it is not a good idea to run any of these machines with the box open. This is because an open cabinet changes the airflow within the filer and often results in reduced airflow to some devices within the box, allowing them to overheat. And that means increased failures. Please, close them up! rgl

1 0

Re: don't run an F630 with the case open
by Christina Sanfilippo 21 May '98

21 May '98

Ouch! That's got to hurt. I am sure Rich was glad he could fix your problem. For the F200/F300 and F500 series filers these filers can run with the case open. As you found out the F630 can't. Cheers, Christina At 08:27 AM 5/21/98 -0400, Tom Yates wrote: >d'oh of the day contest entry, just fyi: > >you can run an F330 with the case open, or at least ajar. i often do >after i've done work, before i seal the case down tight, to make sure that >it boots and i'm not going to have to undo everything to fix it. well, >don't try the same trick with an F630; it doesn't work. thanks to rich >ciammaichella at netapp for not roasting me when we discovered that this >was a major part of my F630 booting problem. > >for those who haven't looked closely inside an F630, there are no longer >cables going from the board to the front-of-case displays; there are now >fixed connectors that seat when the board tray is slid fully home. this >may be true for other high-end models, too - i don't know. > > 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 > > > > > ++++++++++++++++++++****************** Christina Sanfilippo Customer Support Representative Network Appliance 2730 (Shipping:2820) San Tomas Expressway Santa Clara, CA 95051 (408) 367-3747 chrissan(a)netapp.com ============>>>>>>>>>>>>>> Tech Support Hotline:1-888-4-NETAPP (Remember to reference your log # if you have one.) _____________________________________________________ VISIT Netapp On the Web http://now.netapp.com

1 0

don't run an F630 with the case open
by Tom Yates 21 May '98

21 May '98

d'oh of the day contest entry, just fyi: you can run an F330 with the case open, or at least ajar. i often do after i've done work, before i seal the case down tight, to make sure that it boots and i'm not going to have to undo everything to fix it. well, don't try the same trick with an F630; it doesn't work. thanks to rich ciammaichella at netapp for not roasting me when we discovered that this was a major part of my F630 booting problem. for those who haven't looked closely inside an F630, there are no longer cables going from the board to the front-of-case displays; there are now fixed connectors that seat when the board tray is slid fully home. this may be true for other high-end models, too - i don't know. 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

1 0

NTP or rdate.....
by gordon＠nabaus.com.au 20 May '98

20 May '98

Yes, put me on the voting list too. Netapp folk have been misteriously quiet regarding this thread. We know nearly ALL of the netapp dudes read from this list, so how about a response from Netapp? What is Netapps official position regarding NTP? That's my 2cents worth. The check is in the mail.... .... Gordon. Gordon Maudsley, Unix and Network dude, National Australia Bank.

1 0

Re: NetApp File Server (FS230) mounted directly on an NT server
by Keith Brown 19 May '98

19 May '98

Thomas... Taking your various issues one at a time... >It looks like while yes >there is indeed NT server authentication happening now, no matter WHO I >want to access the CIFS shares, they STILL have to MANUALLY PLACED into >the /etc/passwd file!!!! Right. That is the case in current Data ONTAP releases, and the root reason for this requirement is that the security model used by the WAFL file system today is a UNIX security model. In order to assess what a Windows user can and can not do to the files and directories resident on its file system, a filer has to associate each Windows connection (user) with a UNIX identity (a UID and GID(s)) so it can control the users access to data using the UNIX inode based security that it stores (read, write and execute for owning UID, owning GID and other). The /etc/passwd isn't really storing Windows users accounts per se, it's storing mappings between account names and UIDs/GIDs. >What is the POINT of full blown NT server >(domain) authentication if I STILL have to input >ALL of the users into the /etc/passwd file! The Windows clients still get to authenticate themselves in the normal way with the domain infrastructure on startup. The full Windows security/authentication protocols are enforced on the wire, and the users also wind up being subject to all the correct limitations and restrictions that the Windows administrator may have assigned to their account (times of day they are allowed to use the network and so forth...). >We have TONS of >UNIX users across campus who are on various NIS+ domains >and as a result many have overlapping UIDs! So called "overlapping UIDs" situations are not handled by the above mechanism today. If by overlapping UIDs you mean multiple users posessing the same UID, I thought this was generally regarded as a pretty bad idea even to start out with? Unless the accounts with identical UIDs have different lexigraphical user names, that's that. If they do, you could actually get to "break up" the different users with the same UIDs by allocating them different UIDs in the filers /etc/passwd file. However, if you configure the filer into your existing NIS infrastructure as a client, then the problem will remain. As an FYI, we have a utility available on our NOW web site that many of our customers have found useful. Check out CVTUSERS in the "Tools & Utilities" section of NOW. It will automatically generate passwd and group databases from a Windows NT domain infrastructure for you. >Is the ONLY way to get around this is by going to 5.x? Not a current 5.X (5.0) release, no. The resolution to this "problem" as described above is actually quite a bit larger than you might imagine, and is embodied in a major enhancement to the Data ONTAP microkernel and WAFL file system that is already announced and about to be unleashed shortly. Specifically, the solution lies in enhancing WAFL to maintain true Windows NT style security attributes (NTFS Security Descriptors) such that Windows client users can be treated as exactly what they are, Windows client users who might not even have UNIX accounts at all! :-) The new capability is briefly mentioned in the announcement of our previous Windows Network Software enhancement (WNS 2.0) at: http://www.netapp.com/news/level3b/news_rel_971103.html The new software is currently in external test, and has even been briefly reviewed by Windows NT Systems magazine in their article at: http://www.netapp.com/news/level3b/news_rel_971103.html (See the "Sidebar: Midyear Release" box toward the end of the article). I'll save all the details for another time, but suffice to say that what you have coming your way is not just a "resolution", it's a whole new solution in its own right. :-) >Is there any way to MOUNT a NetApp file server directly to an NT file >system and access it like unix does with a /mnt directory? Well, yes. But apparently you already know about it because you write: >I know about the menthod of mounting to NT via a drive letter... but >you can't re-share that as a resource under that the NT file system CAN >you? You mean "re-sharing" a network drive or the directories thereon? No, you can't do that with current versions of Windows NT, at least not without resorting to some undocumented tricks which nobody has ever been able to confirm to me *really* exist. It's like re-exporting NFS mounted file systems on a UNIX machine, arguably not a good idea and definately not allowed. >You know.. like this: > given that > NetApp box="\\barium\C$" > > and > NT Server="\\ntbox" > >net use \\ntbox\shares\share \\barium\share >net use \\ntbox\shares\winapps \\barium\winapps What? Now there's a "net use" syntax that I've never seen before. >Anyone wanna tell me exactly why things can't be done like this? Because it would be a syntax error? :-) Keith

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