toasters

toasters@lists.teaparty.net

13515 discussions

Re: Subscribe
by Brian Tao 19 Feb '98

19 Feb '98

> Date: Wed, 18 Feb 1998 15:11:46 +0000 > From: Lars J�nson <lars(a)datapp.com> > Organization: Data Appliance Scandinavia AB > To: toasters(a)mathworks.com > Subject: Subscribe > > I would like to subscribe > Lars Well, it's not exactly from a *Netapp* employee this time, but it sounds pretty close... ;-) -- Brian Tao (BT300, taob(a)netcom.ca) "Though this be madness, yet there is method in't"

1 0

(no subject)
by Douglas A. Mendoza 19 Feb '98

19 Feb '98

subscribe -- Douglas A. Mendoza Network Appliance, Inc. Website Technical Editor http://now.netapp.com/ phone 408.376.3776 http://www.netapp.com/ e-mail dmendoza(a)netapp.com

9 13

Re: Silly snapshot question
by Daniel Quinlan 19 Feb '98

19 Feb '98

rasmus(a)bellglobal.com writes: > bigdisk> df > Filesystem kbytes used avail capacity Mounted on > / 170788336 71454936 99333400 42% / > /.snapshot 18976480 5180276 13796204 27% /.snapshot > > I was expecting to see 70% utilization on my .snapshot file system > here. Why is it only 27%? From this output I can only conclude that I > have a lot of room and I don't need to bump up my reserve. total disk space 170788336 + 18976480 = 189764816 bytes reserve calculation 189764816 * 10% reservation = 18976480 bytes snapshot usage 5180276 / 18976480 = 27% That's 27% of the amount you reserved. I don't know where you got the 70% figure from. Don't use "snap list" to calculate anything other than which snapshot needs to be deleted first in a crisis. Dan

1 0

AIT vs. DLT7000
by Steve Gremban 19 Feb '98

19 Feb '98

Alexei, We are using DLT7000 (5-10 MB/sec depending on compression) drives which actually outperform the Sony AIT (3-7.8 MB/sec depending on compression). We are not quite seeing 2:1 compression, we get about 60GB compressed on each 35GB tape. -Steve Gremban gremban(a)ti.com ----------------------------------------------------------------------- Previous message follows: I would like to see AIT supported; we currently use DLT, but as our backup needs are approaching 300GB (well, as soon as I get my new filer), DLT is not cutting it (backup/restore times are too long). Also, the media is a huge problem (DLT's are not all too happy with a drop...). I suspect that these additional drives will be supported eventually. Alexei -----------------------------------------------------------------------

3 3

Silly snapshot question
by rasmus＠bellglobal.com 19 Feb '98

19 Feb '98

I was looking at my 630 trying to figure out if it was safe to add another nightly snapshot. According to the Sys-Admin manual on page 10-13 it shows an example where a snap reserve is set to 20% and a 'snap list' shows only 10% utilization. The suggestion is to drop the reserve from 20% down to 12 or 15% assuming usage patterns remain constant. Great, this make sense to me. However, when I look at my system I have: bigdisk> snap reserve Current snapshot reserve is 10% or 18976480 k-bytes. bigdisk> snap list working...................................................... %/used %/total date name ---------- ---------- ------------ -------- 0% ( 0%) 0% ( 0%) Feb 18 12:01 hourly.0 1% ( 0%) 0% ( 0%) Feb 18 08:01 hourly.1 2% ( 1%) 1% ( 0%) Feb 18 00:01 nightly.0 2% ( 1%) 1% ( 0%) Feb 17 20:01 hourly.2 3% ( 1%) 1% ( 0%) Feb 17 16:01 hourly.3 3% ( 1%) 1% ( 0%) Feb 17 12:01 hourly.4 5% ( 3%) 2% ( 1%) Feb 17 08:01 hourly.5 7% ( 4%) 3% ( 1%) Feb 17 00:01 nightly.1 So, so far it would indicate that my 10% reserve is probably a pretty good number for the current setup and if I want to add another nightly snapshot I should probably bump my reserve up a couple of percent to be safe. However, then I do a df: bigdisk> df Filesystem kbytes used avail capacity Mounted on / 170788336 71454936 99333400 42% / /.snapshot 18976480 5180276 13796204 27% /.snapshot I was expecting to see 70% utilization on my .snapshot file system here. Why is it only 27%? From this output I can only conclude that I have a lot of room and I don't need to bump up my reserve. I must have missed something along the way. -Rasmus

3 2

None
by Christina Sanfilippo 18 Feb '98

18 Feb '98

2 1

Subscribe
by Lars J�nson 18 Feb '98

18 Feb '98

I would like to subscribe Lars

1 0

GNU "du" patch to fix .snapshot annoyance
by Daniel Quinlan 18 Feb '98

18 Feb '98

Daniel Quinlan <quinlan(a)transmeta.com> writes: > It's really annoying when find(1) descends into a .snapshot directory. > > Here's a patch against GNU findutils 4.1 tested under Linux and > Solaris to fix the problem. It still allows descent into .snapshot, > but only if .snapshot is specifically listed on the command line. GNU > findutils is available at <URL:ftp://prep.ai.mit.edu/pub/gnu>. Here's the same for "du", which is even more annoying because there isn't a "-prune" option for du. Anyone at NetApp care to comment about whether the inode number for ".snapshot" is constant (and can be counted on)? It always appears to be inode 67. - Dan --- fileutils-3.16/src/du.c.orig Mon Feb 16 00:46:43 1998 +++ fileutils-3.16/src/du.c Mon Feb 16 00:56:35 1998 @@ -54,6 +54,10 @@ #include "save-cwd.h" #include "error.h" +#ifndef S_IRWXUGO +#define S_IRWXUGO (S_IRWXU|S_IRWXG|S_IRWXO) +#endif + #undef convert_blocks #define convert_blocks(b, size) (size == size_kilobytes ? ((b) + 1) / 2 : \ size == size_megabytes ? ((b) + 1024) / 2048 : (b)) @@ -518,7 +522,9 @@ tot_size += size; - if (S_ISDIR (stat_buf.st_mode)) + if (S_ISDIR (stat_buf.st_mode) && + (!(!strcmp(ent, ".snapshot") && (stat_buf.st_uid == 0) && + (stat_buf.st_gid == 0) && ((stat_buf.st_mode & 0777) == S_IRWXUGO)))) { unsigned pathlen; dev_t dir_dev;

2 1

question about multi-RAID
by Daniel Quinlan 17 Feb '98

17 Feb '98

I have a new F630 with 4 shelves (2 pairs) or about 200GB of disk space. Ideally, I would like to configure it with two 100GB RAID groups composed of 12 data drives, one parity, and one spare each (maybe only one spare for the entire filer). If I remake the filesystem with half the disks and start using it, will I be able to upgrade to multi-RAID/multi-volume and add the other half of the disks without a backup/restore cycle? Is there anything else I should know about trying this? - Dan

3 2

Re: Multiple IP Addresses per interface?
by Keith Brown 17 Feb '98

17 Feb '98

From: Brian Tao <taob(a)nbc.netcom.ca> > > According to the man page: > > The interface name vh is used to specify IP virtual host > addresses associated with the filer. Only alias addresses (using > the alias option) may be assigned to the vh interface. > > So I can assign multiple IP's to vh... how do I bind those IP's to >a physical interface? Or is this only used for the virtual hosting >feature in the HTTP server? Unfortunately the doc on this subject doesn't really spell all this out in a manner that could be described as crystal clear, but... My understanding is that the vh interface is just associated with the HTTP daemon's virtual hosting feature. The addresses on vh are the ones that the web server runs through /etc/httpd.hostprefixes to locate a target specific home directory for inbound HTTP requests. However, there is of course the "little" issue of packets destined for a vh address being able to find their way into a filer in the first place, which will inevitably require some cooperation from the real, physical interfaces. This could conceivably be made to happen via routing trickery, perhaps using host routes to ensure that packets destined for vh addresses get fired at a filers physical interfaces on their last hop, or maybe even full network routes, where the vh addresses are on a subnet that is entirely internal to the filer and the physical interfaces form routes to that subnet. Failing this, then you're right, there needs to be some correlation between vh addresses and the addresses being supported by the physical interfaces. Alas I haven't been able to find a description of whether this is a manual or automatic process. I have to assume manual for now. My own experience starts and ends with loading up just the physical interfaces with addresses, and leaving the vh interface alone. Keith

1 0

NetApp 450 end of life
by Mark Richards 17 Feb '98

17 Feb '98

Hello We have a Netapp 450 (the 486 EISA-based machine). NetApp will soon discontinue support for the machine and I'm curious as to what others think of this. We bought our 450 in Nov. 1995 before the PCI-based filers appeared. We are a fairly small shop, (nothing like some of the huge servers discussed on this list :-), but we are delighted with our filer which has never crashed or given us any problems. I understand NetApp reasoning for desupporting the 450 - it costs money to test each new version of the OS on the EISA boxes for what I assume is a small number of remaining users. What do other successful vendors do with older hardware? Does the newest Solaris run on older Sparcs? What about CISCO? I have some seven year old Telebit netblazer routers and they still run the newest OS. NetApp compares its filers to routers, why can't they do what Telebit has been able to do? Our 450 is 27 months old....its performance is terrific for our purposes and I'd have a hard time forming a technical and business case to replace the system with a new model. I guess I just don't think 27 months is VERY OLD! What do you think? I'm not bashing NetApp, they're a great company, I'm just disappointed. Regards Mark Richards Hicks Morley Hamilton Stewart Storie

7 6

None
by Ron Thibault 17 Feb '98

17 Feb '98

subscribe ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Ron Thibault 2820 San Tomas Expressway Network Appliance Santa Clara, CA. 95051 Customer Support Rep. Voice: (408) 367-3742 www.netapp.com Fax: (408) 367-5151 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "The best thing in life is definitely free." - by Big Tent Revival

1 0

None
by Manoj Gandhi 17 Feb '98

17 Feb '98

subscribe ) ========================================================= Manoj Gandhi Customer Support Representative Network Appliance (408)367-3198 http://www.netapp.com/ _________________________________________________________

1 0

Exabyte Mammoth (exb-8900) and 220 changer support?
by Dave Cole 16 Feb '98

16 Feb '98

I am in the process of evaluating tape units and came upon the startling discovery that (as of at least) ontap 4.3R3, there is still no support for the Exabyte Mammoth tape drive. Does anyone know any details about this issue? I would really like to be able to recommend the Exabyte Mammoth 20/40GB tape drives. ------------------------------------------------------------- Dave Cole (DC1110) | dacole(a)netcom.ca Systems Administrator | dacole(a)vex.net | office/~dacole/ Netcom Canada | www.vex.net/~dacole/ 905 King Street West, Toronto, M6K 3G9 | phone - 416.341.5801 Toronto, Ontario, Canada, Earth, Sol | fax - 416.341.5725

1 1

Re: Multiple IP Addresses per interface?
by Keith Brown 16 Feb '98

16 Feb '98

>The purpose of this email is to ask if I can configure multiple IP >addresses per physical interface on an F330. Sure. It's very easy using the "alias" option to ifconfig(1). For example, the sequence: ifconfig e0 192.1.1.1 ifconfig e0 alias 192.1.1.2 will put a couple of addresses on the e0 interface. >If it is possible I did >not see reference to it in the manuals. It's described on the na_ifconfig(1) man page in the command reference guide for Data ONTAP 4.3.X. If you have a NOW account, you can obtain this electronically from: http://now.netapp.com/knowledge/docs/crq41.pdf >the filer is a F330 running 4.0.2c (?) and I have NFS over TCP turned off. Hmmm. I don't know offhand when the multiple IP addresses on a single interface support was formally introduced. There's a chance it was post 4.0.2c, although I'm not certain. If your na_ifconfig(1) man page doesn't mention the "alias" option, you may need to get a hold of a newer software release. Again, you can download that from the NOW web site if you have an account there. Keith

2 1

GNU find patch to fix .snapshot annoyance
by Daniel Quinlan 16 Feb '98

16 Feb '98

It's really annoying when find(1) descends into a .snapshot directory. Here's a patch against GNU findutils 4.1 tested under Linux and Solaris to fix the problem. It still allows descent into .snapshot, but only if .snapshot is specifically listed on the command line. GNU findutils is available at <URL:ftp://prep.ai.mit.edu/pub/gnu>. Dan --- findutils-4.1/find/find.c~ Wed Oct 12 14:21:11 1994 +++ findutils-4.1/find/find.c Sun Feb 15 17:21:29 1998 @@ -38,6 +38,10 @@ #define lstat stat #endif +#ifndef S_IRWXUGO +#define S_IRWXUGO (S_IRWXU|S_IRWXG|S_IRWXO) +#endif + int lstat (); int stat (); @@ -390,8 +394,11 @@ apply_predicate (pathname, &stat_buf, eval_tree); if (stop_at_current_level == false) - /* Scan directory on disk. */ - process_dir (pathname, name, strlen (pathname), &stat_buf, parent); + /* Skip over Network Appliance ".snapshot" directories. */ + if (!(!strcmp(name, ".snapshot") && (stat_buf.st_uid == 0) && + (stat_buf.st_gid == 0) && ((stat_buf.st_mode & 0777) == S_IRWXUGO))) + /* Scan directory on disk. */ + process_dir (pathname, name, strlen (pathname), &stat_buf, parent); if (do_dir_first == false && curdepth >= mindepth) apply_predicate (pathname, &stat_buf, eval_tree);

1 0

Multiple IP Addresses per interface?
by David Drum 15 Feb '98

15 Feb '98

Hello everyone, The purpose of this email is to ask if I can configure multiple IP addresses per physical interface on an F330. If it is possible I did not see reference to it in the manuals. The reason I am asking about this may be of interest to some, though, so an explanation follows. Upon installing a Cisco Catalyst network switch, communication between the filer and our staff machine went bad. I am the only one using the filer from the staff machine so I have only researched the problem as time allows. The staff machine is a Sun Ultra 2 running Solaris 2.5.1, the filer is a F330 running 4.0.2c (?) and I have NFS over TCP turned off. The first thing I did was have our network guys manually configure all the ports on the Catalyst, since this list had discussed the general problem with ethernet autonegotiation. It did not help. The next thing I did was replace the ethernet cable from the staff machine to the Catalyst. If the staff machine was the only device having a problem, it stood to reason. No effect. Next, I began a long series of large packet pings between various machines connected to the Catalyst. As a result of that, I have discovered that a Catalyst does not like multiple subnets and drops from 12 to 22 percent of cross-subnet packets. My network guru claims that the Catalyst switches at layer 2 and so the IP address should not matter, but that certainly is at odds with what I am seeing. I'll let the list know how this turns out OR if you have a Catalyst serving multiple subnets without problem, please let me know how it is configured. Just for the record, the filer and its two main clients are on three successive ports on the Catalyst. 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] You can only measure the size of your head from the inside. [3] When you proceed deliberately, mistakes don't cascade, they instruct. [4] [1] Vernor Vinge [2] Neil Peart [3] Larry Wall [4] Stewart Brand

1 0

[Fwd: Exabyte Mammoth (exb-8900) and 220 changer support?]
by Steve Gremban 13 Feb '98

13 Feb '98

Dave, In the February '98 issue of SysAdmin magazine they do an evaluation of the Sony AIT, Exabyte Mammoth, and the Quantum DLT. They compare head life, tape capacity, cost per GB compressed/uncompressed, media load and file access times, transfer rates compressed/uncompressed, etc., and end up recommending the Sony. I am using Quantum DLT technology and find it performs well, is reliable, and is supported on the Netapp. I'm not sure whether the Sony or Exabyte are supported yet. -Steve Gremban gremban(a)ti.com

2 1

subscribe
by Mark Conley 12 Feb '98

12 Feb '98

1 0

Re: Moving disks between filers
by Robert Hamilton 11 Feb '98

11 Feb '98

Amy, You can install the F220 disks in any available SCSI slot with a SCSI ID between 0 and 6. In the factory default settings on the F540 "shelf pair" these are the remaining 6 slots in the shelf that has the differential converter installed in the far right position. The F220 drives are narrow (8-bit), so they can't be recognized in SCSI IDs >7. A lot of our customers are using recycled F220/F330 disk drives in their newer systems, including the F540, F520 and F630. Regards, Robert Hamilton NetApp >At 10:05 AM 2/11/98 -0500, Amy Chused wrote: >>I have a 540 and a 220. The 220 has been taken out of production and is >>going to be reinitialized before it is used again. The 220 disk shelves >>are full (lots o' disk). >> >>The 540 is in production, and isn't short of disk space right now, but we >>all know how swiftly disk space vanishes. >> >>I'd like to know if it's possible to use disks from the 220 in the 540 disk >>shelves. Does anyone know? >> >>Thanks, >> >> Amy >> >> > >Respectfully, >Greg Collins > > > NETWORK APPLIANCE >========================================================================= > FAST! SIMPLE! RELIABLE! MULTIPROTOCOL!! >========================================================================= >Greg Collins support: 1-888-4NETAPP >District Manager/MidAtlantic support email: support(a)netapp.com >voice: 301-230-5840 main: 1-800-952-5005 >fax: 301-230-5852 URL: http://www.netapp.com/ >email: gcollins(a)netapp.com >pager: 1-800-891-2746 >pager-mail: gcollins-pager(a)netapp.com > >=========================================================================== > > >

1 0

Moving disks between filers
by Amy Chused 11 Feb '98

11 Feb '98

I have a 540 and a 220. The 220 has been taken out of production and is going to be reinitialized before it is used again. The 220 disk shelves are full (lots o' disk). The 540 is in production, and isn't short of disk space right now, but we all know how swiftly disk space vanishes. I'd like to know if it's possible to use disks from the 220 in the 540 disk shelves. Does anyone know? Thanks, Amy

1 0

Logging in to CIFS from W95 using multiple loginids
by Steve Losen 10 Feb '98

10 Feb '98

One of our staff is working on an application that makes it easy to map a netapp share to a drive letter. This application works fine on NT, but W95 is causing trouble. The application is a little visual basic program that opens up a box that allows the user to enter a loginid, password, and shows a list of drive letters to map to. The user sets these things and the application supplies the name of the share by building it from our NetApp's name and the user's loginid, eg, \\home1\abc9d. On NT this works great. You can specify different loginids and passwords for different shares and map them to various drive letters. On W95, however, this doesn't work so well. Apparently the calls that the visual basic app makes are replacing the loginid that the user specifies with the loginid that the user logged in to W95 with. In other words, this app cannot override the current W95 loginid. If you specify the same loginid that you logged in to W95, then it works fine. We are curious if anyone else has any experience with this problem and has any suggestions for fixing it. For the curious, we want an application like this because our NetApp box is going to eventually have around 20,000 CIFS accounts and we are going to use the unix passwd authentication rather than set up a NT domain controller and have to pay license fees per user, etc. We don't plan to have any WINS servers know about the NetApp. We have too many different networks at this University to rely on every net being able to browse for the NetApp in Network Neighborhood, etc. Instead, we are going to hand out this little app that knows the name of the netapp, making browsing, and/or remembering the name of the server unnecessary. Plus, this app will disconnect after a certain amount of idle time. Because various departments maintain their own NT file servers (with their own loginids), we can't assume that a user will always be able to login to W95 using the same loginid that they have on the netapp, hence the need to override the W95 loginid. Also, in some of our lab environments, each PC automatically logs in to W95 on boot using a "dummy" machine loginid. Has anyone done anything like this? If we can't override the W95 loginid, could our application somehow change it, connect to the server, and change it back? Steve Losen scl(a)virginia.edu phone: 804-924-0640 University of Virginia ITC Unix Support

1 0

Re: More capacity on lower-end filers
by hitz＠netapp.com 10 Feb '98

10 Feb '98

> On Thu, 12 Jun 1997, Dave Hitz wrote: > > > > Although marketing certainly has input, the space restrictions also > > come from engineering droids. During normal operation the smaller > > machines could certainly handle more disk, but the time required for > > something like RAID reconstruction could get dangerously long. > > Is this the reason why the F210 can still only use around 50GB of > disk, even though it can physically accomodate more using 9GB drives? In order for a RAID reconstruction to complete, you have to read all of the data on all of the other disks. So RAID reconstruction time is proportional the the amount of data, not proportional to the number of disks. Dave Hitz hitz(a)netapp.com Network Appliance (408) 367-3106

8 10

Unexpected behaviour from F330 running 4.0.1c and Solaris 2.5.1
by Christoph Doerbeck 09 Feb '98

09 Feb '98

Greetings, we've isolated our little problem but... the behaviour or the solaris system seemed a bit unexpected to me. There exhisted a >2GB file on our NAS and when the Solaris box encountered it, all communication (NFS) to the toaster became locked. We could ping, telnet etc... but no NFS. Once we tracked down the culprit, removing the file promptly returned the system to normal. If I understand things correctly, Solaris 2.6 now allows larger files >4GB. Anyway, I thought I'd post as a informational blurb, and maybe someone could outline the latest file (system) limits for the latest versions of the popular operating systems (HP,SGI,AIX, ..er...cough....gag...NT, and others) Maybe this would be a good QA for the NetApp FAQ some of you discussed a few months ago. - Christoph

1 0

archive is now searchable... ish.
by Tom Yates 09 Feb '98

09 Feb '98

The archive can now be searched through a simple search engine. Note that this search engine, like the web server itself, runs on my desktop machine. If I find my work going very slowly because you're all searching the archives like crazed weasels, I'm likely to kick back. Please be nice. The search engine can be found at http://teaparty.mathworks.com:1999/toasters/Search.html . In case anyone wasn't aware, the list is archived at http://teaparty.mathworks.com:1999/toasters/ . There is a reference to the search engine in the "about this archive" page, which is referenced at the very very top of the main archive page, in case you forget where the engine is. If people find this search engine not working, or think it could be improved in ways that won't take me a week, please let me 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

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