toasters

toasters@lists.teaparty.net

13515 discussions

Re: Network Appliance makes the news...
by Jim McCoy 18 May '99

18 May '99

sirbruce(a)ix.netcom.com writes: >As to what happened, the most *likely* scenario for any signficant >downtime of the Netapp is double disk failure. That is, one disk >failed, and during reconstruction they lost another. I beg to differ. YahooMail has had significant problems with all sorts of unpleasant little bugs which has led to downtime for the system and in two cases actual data loss. Since the beginning of this year we have apparently run into every possible failure mode of these systems, except for a dual disk failure, and I can assure you that there are all sorts of reasons that an event like this could occur. Reading the tea leaves on this one I would guess that they also run into bug 1549 and were forced to run a wack on a filer (props to John E. for the fast wack, if it were not for that single improvement I can assure everyone at NetApp reading this that Yahoo would be using EMC boxes right now...) >If they were running a cluster, most likely there would have been >no interruption in service. You get what you pay for. I'm sure >their Apache (or whatever) web servers have crashed more than their >filers have. How I wish we would get what we paid for... We run a hacked version of Apache and have probably had a total of 30 minutes of system downtime due to this Apache problems this year (mostly due to a memory leak I think that we introduced into our version.) Contrast that with a total of 36 hours of system downtime since January due to filer failures. If NetApp expects us to pay twice as much per MB of storage and run a cluster then they are going to start getting into the realm where an EMC box begins to look like a better deal, and when it comes to system reliability NetApp still has a long way to go to reach the sort of reputation that EMC has in the industry. NetApp had a very serious problem with quality control on hardware and QA within the engineering department was so useless as to be non-existent. We have been given assurances that this specific problem as been recognized and is being addressed. Only time will tell. >There doesn't seem to be much of a story here. But there is a story here, and one which needs to be addressed. NetApp charges a large premium over the raw costs of the disks and hardware to hook them together compared to what I could rig up with a PC and a few raid controllers, as a customer I would like to know that I am getting some value for this addtional cost. When that "value-add" proves to be a mirage I think that I have a justifiable reason to be pissed off. Critical Path was just the first customer to admit that NetApp was the source of the downtime, the prompt attention we have received from NetApp (even when there was nothing they could do) is the only thing which prevented Yahoo from being the first one to attribute responsibility for system downtime to NetApp in a press release. jim mccoy Yahoo! Inc.

5 4

Re: Network Appliance makes the news...
by Tripp Cox 18 May '99

18 May '99

On 17 May 1999 20:54:25 -0400, Andrew Hobson wrote: >Of course, I'm more Netapps means more hardware, and more opportunity >for foulups, but we've received cards that were bent so badly we >couldn't put them in. The box didn't appear damaged at all. I recall several instances of buggy hardware and sloppy hardware QA dating back to early-mid 1997. I'd like to think things have improved since then. We had to replace all of our SCSI controllers in a slew of F540s, we had bad EMUs in a couple of F630s, and a heck of a time with the short-lived EFBs of the F520s. Not to mention a plethora of bad disks (new equipment) and shelf pairs bent beyond any utility. I also recall having to reseat a small microchip within the casing of an EFB. Of course, Jan and I managed to hose Netapp's lab equipment so bad on one trip to San Jose that a stream of curious engineers visiting the sick equipment couldn't recover it before the day was over. Maybe we're not the best test case. :) Of course the point to all of this being that it generated a great deal of service downtime that we never should have had to endure. Tripp

1 0

Re: Network Appliance makes the news...
by Andrew Hobson 18 May '99

18 May '99

On Mon, 17 May 1999 12:50:58 -0700, "Jim McCoy" <mccoy(a)yahoo-inc.com> said: > sirbruce(a)ix.netcom.com writes: >> As to what happened, the most *likely* scenario for any signficant >> downtime of the Netapp is double disk failure. That is, one disk >> failed, and during reconstruction they lost another. > I beg to differ. I just about to write this exact same email. We've had the same experiences with filer failers and quality control. We're having the same thoughts about the cost when you want non-stop computing. Netapp is great for the 98% uptime market. They aren't ready to play in the 99.99% uptime market, at least in my view. I used to be a Netapp enthusiast, but no longer. I'd sleep better if our customers' email was moved off of filers. Drew "Speaking from experience, but not speaking as a representative of MindSpring" Hobson -- "He's a love machine, sir." -- Waylon Smithers

2 2

Re: Network Appliance makes the news...
by Jeff Newton 18 May '99

18 May '99

Ok, I'll bite. What is a "wack"? Cheers, > In the immortal words of Jim McCoy (mccoy(a)yahoo-inc.com): > > > Reading the tea leaves on this one I would > > guess that they also run into bug 1549 and were forced to run a wack on a > > filer (props to John E. for the fast wack, if it were not for that single > > improvement I can assure everyone at NetApp reading this that Yahoo would be > > using EMC boxes right now...) > > Out of morbid curiosity, which is bug 1549? I have had to run wack > in recent memory, and it was a...sobering experience. > > -n > > ------------------------------------------------------<memory(a)blank.or g> > I've got more than one membership / to more than one club > and I owe my life / to the people that I love. (--Ani DiFranco) > <http://www.blank.org/memory/>---------------------------------------- -- ---- Jeff Newton Unix Systems Administrator PMC-Sierra Inc.

2 1

32-bit uptime counter?
by Brian Tao 17 May '99

17 May '99

Still running good ol' 4.2a on an F230: Sun Apr 18 21:00:00 EDT [statd]: 9:00pm up 497 days, 1:10 407777519 NFS ops, 0 CIFS ops, 0 HTTP ops Sun Apr 18 22:00:00 EDT [statd]: 10:00pm up 497 days, 2:10 408458997 NFS ops, 0 CIFS ops, 0 HTTP ops Sun Apr 18 23:00:00 EDT [statd]: 11:00pm up 42 mins, 409036765 NFS ops, 0 CIFS ops, 0 HTTP ops Mon Apr 19 00:00:00 EDT [statd]: 12:00am up 1:42 409494263 NFS ops, 0 CIFS ops, 0 HTTP ops This looks like 2^32 ticks elapsed (about 42949673 seconds)... is this one of the counters being increased to 64 bits? -- Brian Tao (BT300, taob(a)risc.org) "Though this be madness, yet there is method in't"

2 1

Re: Running CIFS
by sirbruce＠ix.netcom.com 17 May '99

17 May '99

On 05/05/99 15:53:45 you wrote: > >In message <7F608EC0BDE6D111B53A00805FA7F7DA03A83938(a)TAHOE.netapp.com>, "Muh >lestein, Mark" writes: >>A mixed qtree already works exactly as you want it to. In a mixed qtree the >>security of a file is based on the last security-setting operation. If an A >>CL is >>set which denies you access, but you still need to do something with a file, >>then as root you can change the security on any files/dirs in that qtree (using >>chmod, chown, or chgrp), which changes the security-style to UNIX. After that, > > Ok, so far I've read this alot. No maybe I'm missing something > 'cause I get the following errors trying to put it into practice. What errors? These are exactly correct. > Here's a test on a directory w/ NTFS permissions allowing only full > control to the owner: In other words, you've told NTFS that Administrator can't read the directory or change permissions. All Administrator can do is assume ownership. >root@jagular: fs0/home/kelleher [130] # >root@jagular: fs0/home/kelleher [130] # ls -ald test_dir >drwx------ 2 kelleher users 4096 May 5 14:47 test_dir >root@jagular: fs0/home/kelleher [131] # >root@jagular: fs0/home/kelleher [131] # ls -al test_dir >test_dir: Permission denied >total 8 This is correct; root on UNIX will map as Administrator in the NT realm, so it can't read the directory. >root@jagular: fs0/home/kelleher [132] # >root@jagular: fs0/home/kelleher [132] # chown 755 test_dir >chown: test_dir: Not owner Firstly, you should have done chmod, not chown. Secondly, this would be correct... Administrator (root) does not have permission to change the file permissions. >root@jagular: fs0/home/kelleher [133] # >root@jagular: fs0/home/kelleher [133] # chown root test_dir >chown: test_dir: Not owner >root@jagular: fs0/home/kelleher [134] # This should have suceeded, if I understand the model correctly. I'm not sure why it didn't; perhaps it's a bug? > I can however su to the UNIX user, chmod it, and then get access to > the file: > >root@jagular: fs0/home/kelleher [134] # su kelleher >jagular% >jagular% chmod 755 test_dir >jagular% >jagular% exit >jagular% root@jagular: fs0/home/kelleher [135] # >root@jagular: fs0/home/kelleher [135] # ls -al test_dir >total 72 >drwxr-xr-x 2 kelleher users 4096 May 5 14:47 . >drwxr-xr-x 25 kelleher users 32768 May 5 15:25 .. >-rwx------ 1 kelleher users 0 May 5 14:47 test_file >root@jagular: fs0/home/kelleher [136] # >root@jagular: fs0/home/kelleher [136] # > > Is this what you mean? This all makes sense. >>those files follow regular UNIX rules, which allows root full access. If the >>user wants to put an ACL back on afterwards, no problem. >> >>We have found that using ACLs in a mixed qtree can be helpful in a number of >>situations. That's because ACLs are enforced even for NT Admins and root. For >>example, we found that some files were being accidentally deleted by root/Admin >>users, so we put an ACL on those files which allows only READ access to all >>users. That prevents deleting, even by root or NT Admins. Of course, if we >>actually want to delete the files we just change the permissions to allow that, >>since in a mixed qtree both root and NT Admins have a special dispensation >>tochange the permissions. > > Sounds like the problem is too many people with root/Admin access. > Your solution only requires these people to take an extra step > (give themselves the permissions) before they accidentally mess > something up. If I were you I'd work on getting rid of their > access. (Not that it's an easy thing to do.) In my experience many people are unwilling to take this approach, for reasons I can't quite fathom. They would much rather let incompetent people continue to work for them and spend effort trying to keep them from shooting themselves in the foot rather than spending the time to make them competent and/or getting another employee. Bruce

2 1

ISP_2100: cmd opcode 0x28, retry count 0: command timeout, aborting request error
by Mike Ball 17 May '99

17 May '99

Can anyone shed some light on the error below? It looks like it is just a media error which was recovered from, but I wanted verification. Thank you, Mike Wed May 12 00:13:03 EDT [isp2100_timeout]: 8.45 (cmd opcode 0x28, retry count 0): command timeout, aborting request Wed May 12 00:13:03 EDT [isp2100_timeout]: isp2100_reset_device: device 8.45 (1.45)

1 0

determining patch revision level..
by Kane, Kevin 17 May '99

17 May '99

Can anyone tell me how to determine the patch level of the OS? Does a response of 5.2.1 from the "version" command mean that I have generic 5.2.1 with no P releases? TIA Kevin A. Kane Manager - Engineering Computing Infrastructure Aspect Telecommunications Direct: 978 250-7933 Business Fax: 978 250-7998 Pager: 888 799-4866 Home: 978 388-1875 Mobile: 508 783-2695 Email: Kevin.Kane(a)aspect.com Pager Email: Kevin.Kane.Pager(a)aspect.com

4 3

Re: determining patch revision level..
by Robert Palermo 17 May '99

17 May '99

Yes you are running the generic 5.2.1 as it were. If you had a patch release the console command "version" would report 5.2.1P2 for example. There are other trailing characters such as D8 and such but the system should report the actual patch version if you are running it. > From: "Kane, Kevin" <Kevin.Kane(a)Aspect.com> > To: "'toasters(a)mathworks.com'" <toasters(a)mathworks.com> > Subject: determining patch revision level.. > Date: Mon, 17 May 1999 12:51:18 -0400 > MIME-Version: 1.0 > > Can anyone tell me how to determine the patch level of the OS? Does a > response of 5.2.1 from the "version" command mean that I have generic 5.2.1 > with no P releases? > > TIA > > Kevin A. Kane > Manager - Engineering Computing Infrastructure > Aspect Telecommunications > Direct: 978 250-7933 > Business Fax: 978 250-7998 > Pager: 888 799-4866 > Home: 978 388-1875 > Mobile: 508 783-2695 > Email: Kevin.Kane(a)aspect.com > Pager Email: Kevin.Kane.Pager(a)aspect.com > >

1 0

Re: Ontap P* relase mess.
by Jason D. Kelleher 17 May '99

17 May '99

In message <Pine.BSF.3.96.990506145048.2243D-100000@localhost>, kg writes: > >So it looks like I'm going to have to downgrade to 5.2.1P2 from 5.2.2 to >avoid panics from assertions in the netbios name service.. Patches for >5.2.2 are somewhere in the indeterminate future.. (all of this for bug >12712, which for some reason isn't listed in Bugs Online) > >This brings up one of the points that Tom (?) was ranting on a week or so >ago.. Why release 5.2.2, a minor bug-patch revision, that don't >incorporate patches from its predecessor? > >Perhaps I shoud've called support first, but I just brought these 760's >online with 5.2.2, under the apparently false impression that I was >getting all the up-to-date fixes.. Was there an erratta for 5.2.2 that I >missed that said "THE FOLLOWING IS STILL BROKEN:"? > >argh. > >..kg.. Sounds like another vote for a concise "Known Bugs" list for each ONTAP rev... jason

1 0

Re: Network Appliance makes the news...
by sirbruce＠ix.netcom.com 17 May '99

17 May '99

On 05/14/99 23:13:56 you wrote: > >On Fri, May 14, 1999 at 10:28:43PM +0000, Richard L. Rhodes wrote: >> Our EMC sales rep sent this to me. Any thoughts or comments, >> on what might have happened? >> >Sure. I know exactly what happened. > >Your EMC rep peed his pants when he saw a competitor having trouble >in public, and made sure you were aware of it. Pretty typical of EMC >reps, AFAICT. They also went and posted it on the NTAP Yahoo message board. As to what happened, the most *likely* scenario for any signficant downtime of the Netapp is double disk failure. That is, one disk failed, and during reconstruction they lost another. However, the description of events (moving mailboxes) may or may not support this. It's hard to rely on such technical details in a media report. It's not clear whether or not data was actually lost. It's also possible that the had a crash due to a bug and this left the filesystem is an inconsitent state and/or a state that caused repeated crashes. In both this and the double-disk failure case, they would have to run wack to make sure the filesystem was okay before returning it to service. This sounds like the "cleaning" they are talking about. If they were running a cluster, most likely there would have been no interruption in service. You get what you pay for. I'm sure their Apache (or whatever) web servers have crashed more than their filers have. There doesn't seem to be much of a story here. Bruce

3 2

Network Appliance makes the news...
by Richard L. Rhodes 15 May '99

15 May '99

Our EMC sales rep sent this to me. Any thoughts or comments, on what might have happened? THanks rick ======================================================================== Failure hits Critical Path email By John Borland <mailto:jborland@cnet.com> Staff Writer, CNET News.com May 11, 1999, 6:05 p.m. PT An equipment failure at email outsourcing company Critical Path has kept potentially tens of thousands of email and Web hosting customers out of their mailboxes since yesterday morning. San Francisco-based Critical Path <http://www.cp.net>, which provides Web-based email services for companies such as America Online <http://www.aol.com>, US West <http://www.uswest.com>, and Sprint <http://www.sprint.com>, said one of its three core mail storage devices failed yesterday morning. The company would not comment on how many customers or email addresses were left inaccessible. One of the company's clients, Verio.com's <http://www.verio.com> TABNet Web hosting subsidiary, has between 15,000 and 20,000 customers, most of which rely on the Critical Path systems, a Verio spokesman said. According to Critical Path, the outage was sparked by the failure of a Network Appliance mail storage system. All the mailboxes affected were moved to another machine, but the data on the failed machine had to be "cleaned" before service could be restored, a spokeswoman said. Network Appliance is doing the equipment repair and data recovery work, and the companies hope to have the mailboxes back online by 10 p.m. today, the spokeswoman said. Critical Path said any email sent to one of the affected addresses is being kept in storage and will be delivered to the unavailable mailboxes as soon as service is restored. In an apparently unrelated outage, many Netcom ISP subscribers also found themselves without access to email beginning early yesterday. The company, which was recently acquired by MindSpring <http://www.mindspring.com>, said that one of its mail servers failed, largely affecting customers with last names beginning with "D," according to MindSpring spokesman Ed Hansen. Other customers also were stuck with slower or temporarily unavailable service as the volume of mail traffic outpaced Netcom's remaining server capacity, he added. The company said it expects to have service restored by the end of the business day today, Hansen said. Customers with affected mailboxes will likely see a deluge of mail, followed by a slower trickle of backlogged messages as they are re-created from the failed server, he added. ---------------------------------------------------------------------- Richard L. Rhodes e: rhodesr(a)firstenergycorp.com Ohio Edison Co. p: 330-384-4904 f: 330-384-2514

2 1

Re: 100% Cpu using quotas under 5.x
by sirbruce＠ix.netcom.com 14 May '99

14 May '99

On 05/14/99 01:10:13 you wrote: > It's the [de0|de1|hipri|lopri]-bit that puzzles me a bit - > can anyone explain ? I can't address your general problem (as you've said in another post, it appears to be SNMP-related) but as to this I've been told that this is just the particular thread the software happened to be in when the quota was to be exceeded. Seeing this in the wafl layers makes sense; I'm not exactly sure how this can occur in the network (de) modules but I've seen it before and been told there's no problem here. I agree, it is a little confusing, but I suppose it's better than not getting the information at all. :) Bruce

2 1

Stale NFS errors
by Jay Orr 14 May '99

14 May '99

HI I'm sure this is a frequently asked question, but I'm stumped... Running an F330 (4.2a) on some sun 2.5.1 systems. /var/adm/messages gives : May 12 23:51:52 cleats unix: NFS write error on host hardrock: Stale NFS file handle. May 12 23:51:52 cleats unix: (file handle: 939e1400 577cd30c 0 856f1000 49ee7326 0 2ae0500 51356400) May 13 00:07:24 cleats unix: NFS write error on host hardrock: Stale NFS file handle. May 13 00:07:24 cleats unix: (file handle: 939e1400 577cd30c 0 816f1000 3fda490e 0 2ae0500 51356400) May 13 03:13:35 birdland unix: NFS write error on host hardrock: Stale NFS file handle. May 13 03:13:35 birdland unix: (file handle: 939e1400 577cd30c 0 856f1000 d1f07326 0 2ae0500 51356400) May 13 03:24:07 therat unix: NFS write error on host hardrock: Stale NFS file handle. May 13 03:24:07 therat unix: (file handle: 939e1400 577cd30c 0 866f1000 9b927326 0 2ae0500 51356400) May 13 11:38:28 mcbutts unix: NFS write error on host hardrock: Stale NFS file handle. May 13 11:38:28 mcbutts unix: (file handle: 939e1400 577cd30c 0 866f1000 df967326 0 2ae0500 51356400) Now, I'm trying to figure out WHAT file(s) are doing this. I've read NetApps "Troubleshooting NFS Stale File Handles on Sun Clients (Series of FAQs)" doc, and SunWorld's "Errno Libretto" which discuss tracing these things, but nothing is what it should be. **1) "use showfh". That's not in 2.5.1. I got a copy of fhfind, but that doesn't work either. 1a) First, it looks in /etc/mnttab for the device id to decide where to search. NO DEVICE ID MATCHES!!!! 1b) it uses the 4th number for the inode number. WRONG -- that translates into a 10 digit INODE number. All the inode numbers are 5-7 digits. **2) NetApp's doc states that the second number is the device number of the file system, and the third number is the inode number. NO, the 3rd number is "0". My guess is that the filer uses a 64bit INODE number, while solaris is using a 32bit inode. How can I figure out what is going on?? ----------- Fujitsu - Nexion, St. Louis, MO Jay Orr (314) 579-6517

3 8

RE: Multiple subnets, NT and Netapp
by Mankovsky, Mike 14 May '99

14 May '99

Give this a try this. Use -wins option with the ifconfig command on filer to suppers the registration of both filer IP addresses with WINS. Use OPTIONS CIFS.NETBIOS_ALIASES to add a second netbios name for each filer that corresponds to the filer's second NIC. Add a static entry to WINS that maps the first netbios name to the ip address of the first NIC. Do not make this a multihomed entry. Add a second static entry to WINS that maps the netbios alias name to the ip address of the second NIC. Do not make this a multihomed entry. >From the IIS machine use UNC path to connect to the filer's alias name. Mike Mankovsky Network Appliance -----Original Message----- From: Sam Schorr [mailto:sschorr@homestead.com] Sent: Friday, May 14, 1999 8:47 AM To: 'toasters(a)mathworks.com' Subject: Multiple subnets, NT and Netapp Folks, I hoping someone out there can help us understand how we can setup a network "backchannel" in an NT domain environment. Here's the situation: One NT domain that has approximately 20 servers in it. Of those, 8 are IIS servers. The HTML content, dynamically created and accessed by anonymous web users, is stored on 2 F740's. Right now, all the servers, domain controllers, SQL servers, and the filers are on one subnet. The filers are accessed by dlls on the IIS systems using UNC - \\filername\dir1\dir2\websitename. The filername part of that path conforms to the hostname given the filer in the NT domain; the websitename part if dynamically created when the anonymous web user creates a web site. What I want to do is have the filers connect to the IIS systems, and only the IIS systems, on a second subnet to maximize bandwidth and isolate the file access traffic from other traffic. However, I cannot seem to get around the fact that NT wants to take the UNC pathing and do a name resolution through WINS; and WINS does not like 2 network addresses for the same host. NT does not seem to handle the idea that allows name resolution by NIC, something easy to setup on the filer and in UNIX. When we have tried this, the IIS systems seem to ignore lmhost files where we have put the second filer NIC name and address, routes a call to WINS, then finally seems to find the second subnet when WINS fails to. We have tried putting the IP address of the second NIC in the UNC call, but NT treats it as a string, so no go. BTW, the IIS servers do have a second NIC card on the backchannel switch that we connect the second filer NIC to. Help? Sam Schorr Homestead Technologies ph: (650) 549-3152 fax: (650) 364-7329 sschorr(a)homestead.com Don't just build a home page...build a Homestead!! http://www.homestead.com

1 0

Multiple subnets, NT and Netapp
by Sam Schorr 14 May '99

14 May '99

Folks, I hoping someone out there can help us understand how we can setup a network "backchannel" in an NT domain environment. Here's the situation: One NT domain that has approximately 20 servers in it. Of those, 8 are IIS servers. The HTML content, dynamically created and accessed by anonymous web users, is stored on 2 F740's. Right now, all the servers, domain controllers, SQL servers, and the filers are on one subnet. The filers are accessed by dlls on the IIS systems using UNC - \\filername\dir1\dir2\websitename. The filername part of that path conforms to the hostname given the filer in the NT domain; the websitename part if dynamically created when the anonymous web user creates a web site. What I want to do is have the filers connect to the IIS systems, and only the IIS systems, on a second subnet to maximize bandwidth and isolate the file access traffic from other traffic. However, I cannot seem to get around the fact that NT wants to take the UNC pathing and do a name resolution through WINS; and WINS does not like 2 network addresses for the same host. NT does not seem to handle the idea that allows name resolution by NIC, something easy to setup on the filer and in UNIX. When we have tried this, the IIS systems seem to ignore lmhost files where we have put the second filer NIC name and address, routes a call to WINS, then finally seems to find the second subnet when WINS fails to. We have tried putting the IP address of the second NIC in the UNC call, but NT treats it as a string, so no go. BTW, the IIS servers do have a second NIC card on the backchannel switch that we connect the second filer NIC to. Help? Sam Schorr Homestead Technologies ph: (650) 549-3152 fax: (650) 364-7329 sschorr(a)homestead.com Don't just build a home page...build a Homestead!! http://www.homestead.com

1 0

Re: 100% Cpu using quotas under 5.x
by Ketil Kirkerud Elgethun 14 May '99

14 May '99

Following up on my own posting: It looks a _lot_ like SNMP is the culprit - we have a script, run daily, that extracts information (via SNMP) from the filer about qtrees. Running this scipt by hand made the filer go bananas (CPU pegged at 100%, until i did a 'quota off;quota on'. The interesting bit of the script is as follows : for $host (ListService("netapp-server", "ALL")) { mkdir($host, 02775) unless (-d $host); # netapp.netapp1.quota.quotaState.0 ($state) = snmpget ($host, $cmty, '1.3.6.1.4.1.789.1.4.1.0'); if ($state != 2) { warn "$host does not have quotas initialized (state = $state)"; next; } $id=0; while (1) { $id++; # netapp.netapp1.quota.qrTable.qrEntry.qrIndex # netapp.netapp1.quota.qrTable.qrEntry.qrType @ret = snmpget ($host, $cmty, '1.3.6.1.4.1.789.1.4.3.1.1.'.$id, '1.3.6.1.4.1.789.1.4.3.1.2.'.$id, '1.3.6.1.4.1.789.1.4.3.1.8.'.$id, ); last if $ret[0] < 0; next if $ret[1] != 3; # Only qrTypeTree $ret[2] =~ y/A-Z/a-z/; ($fs = $ret[2]) =~ s%/%-%g; print CFG <<EOF; << print to mrtg.cfg-file >> EOF; } } So. This works fine on 4.x - and it looks like it works fine on a F330 running 5.1pre (but this box does not have 160K users on it). Anyone have any ideas ? -- ---Ketil Kirkerud Elgethun, SOL System

1 0

100% Cpu using quotas under 5.x
by Ketil Kirkerud Elgethun 13 May '99

13 May '99

I have an interesting problem with one of my F540s (maxed out on memory and disk controllers, approx 40 disks) : It used to run 4.3.1D3 - on recommendation from my vendor (there was a bug in the release we were running that bit us), I upgraded to 5.2.1P2D4 - the NetApp is serving the home directories of approx. 160.000 users (largest UID approx 161000)- and we use default quotas. A few (less than 500) of the users have specific quotas. This worked very well under 4.x - it still works, but after having been up and running for 20 hours or so, the NetApp becomes _very_ sluggish, with 100% CPU consumed - it still serves nfs requests, but very slowly. Doing 'quota off; quota on' gets the box going again - but this is hardly ideal. I can imagine some mechanisms that would cause this: 1) Bugs have crept into the quota software in OnTap 2) Bugs have crept into syslog - we run with the default syslog config, and we get a _lot_ of messages about exceeded quotas. The messages differ slightly : Thu May 13 21:26:12 MES [wafl_hipri]: uid 75355 tid 15: disk quota exceeded on volume vol0 Thu May 13 21:27:31 MES [de1]: uid 75355 tid 15: disk quota exceeded on volume vol0 Thu May 13 22:26:44 MES [de0]: uid 18002 tid 16: disk quota exceeded on volume vol0 Thu May 13 22:26:44 MES [wafl_lopri]: uid 18002 tid 16: disk quota exceeded on voume vol0 (these are just excerpts, we get a _lot_ of them as well as quite a few "Thu May 13 21:40:59 MES last message repeated 7 times" related to the aforementioned messages) It's the [de0|de1|hipri|lopri]-bit that puzzles me a bit - can anyone explain ? I couldn't care less about users exceeding their quotas - and wouldn't mind being able to turn off the reporting. (I would still like to be able to know about qtrees exceeding their quotas, however). My gut feeling is that this is the root of the problem. (too many 'quota exceeded' messages overloading syslogd somehow, that is) 3) The quota mechanisms can't take the load (but then, why did it work so well in 4.x ?) Has anyone else seen tha same problem ? (It's being escalated through our vendor - my experience with them and NetApp is good, but I'd like to find out if anyone else has seen anything like this problem) -- ---Ketil Kirkerud Elgethun, SOL System

1 0

RE: Netapp filers & MS Exchange
by Jonas Jonsson 13 May '99

13 May '99

I've tried to do some "brain work" on this and come to the following conclution: 1. How can I create a "Exchange consistent" snapshots to provide for the backup of Exchange (snapshot) with of the shelf software like Legato Networker (no Exchange API - just files). Snapshtos will do for this. Just remember Exchange might be running for a while in a non-consistant mode before it decides to do a halt. 3. In case of a need for partial data recovery - is there a Filer local copy, the convert a part of a snapshot from RO to RW mode besides a through the host copy can can always be done but consumes a lot of bandwidth, time and resources. Would appreciate any thoughts/suggestions/enlightnings/pointers to the above subjects. Partial recovery I think might be a problem which ever solution you chose. Exchenge, as far as I know, is working on a information blob. The best solution is to use snapshots and recover to a stable point in time, preferably some hours back in time. If anyone has information on the workings/non-workings of Exchange and how it stores/not stores information. Pleas comment or I'll go further on my darkend path ;) /// Regards, Jonas

1 0

Re: Netapp filers & MS Exchange
by sirbruce＠ix.netcom.com 13 May '99

13 May '99

On 05/13/99 07:30:54 you wrote: > >I am willing to implement Exchange on NetApp. I have two exchange >servers with each holding about 20GB of data. I have seen the document >on NOW. It is imperative to have close to 24*7 service from the Exchange >servers. You can't, because you've already chosen Exchange. There's nothing wrong with Netapp in this case, but Exchange is a crappy piece of software that may give you 24*6 service at best. Bruce

1 0

Re: Syslog message
by sirbruce＠ix.netcom.com 13 May '99

13 May '99

On 05/13/99 10:24:30 you wrote: > >Jay Orr wrote: >> >> One last question on this topic -- I had removed the syslog.conf file from >> the toasters yesterday to stop the messages, but it kept logging anyway. >> If it was a unix box, I'd just send a HUP signal to the syslogd, but how >> does a toaster know the file has changed and needs to reconfigure (or that >> the file is missing)? > >It checks the file every so often (every 30 seconds, IIRC), and updates >automagically. And, just the clarify, if you remove the file entirely I think it just falls back to the "default" configuration. It would take an empty syslog file to eliminate most messages (but I think some would still get logged anyway). Bruce

1 0

change notification handles
by Todd C. Merrill 13 May '99

13 May '99

Some users here have recently reported significant slowdowns with some testing suites. They claim the addition of a few directories to a certain path triggered the performance degradation and are asking me about change notification handles for the CIFS service. Searching the NOW site, CD docs, and the toaster mailing list archives just gave me hits on "Yup, it's there," but nothing else. Can anyone reference a pointer on more information? What the ONTAP limit is? If there are any tunable parameters? NetApp Release 5.2.1P2: Tue Mar 16 10:41:03 PST 1999 Until next time... Todd C. Merrill The Mathworks, Inc. 508-647-7792 24 Prime Park Way, Natick, MA 01760-1500 508-647-7012 FAX tmerrill(a)mathworks.com http://www.mathworks.com ---

2 1

Syslog message
by Jay Orr 13 May '99

13 May '99

Anyone seen this or know what it means? I just setup the syslog to send to another system (it didn't have a syslog.conf file before). I set it up like the example so it's sending *.err and kern.* to another system. Now I'm getting : May 12 11:57:31 hardrock Wed May 12 11:57:32 CDT [portmapper]: portmap: prog (300214,1,201) not registered. May 12 11:57:31 obriens Wed May 12 11:57:32 CDT [portmapper]: portmap: prog (300214,1,201) not registered. in the syslog (hardrock and obriens are the two toasters). It's repeating every half hour or so... Thanks in advance for the help... ----------- Fujitsu - Nexion, St. Louis, MO Jay Orr (314) 579-6517

2 5

Netapp filers & MS Exchange
by Itzik (Itzhak) Meirson DSc 13 May '99

13 May '99

I am willing to implement Exchange on NetApp. I have two exchange servers with each holding about 20GB of data. I have seen the document on NOW. It is imperative to have close to 24*7 service from the Exchange servers. My questions are related to backups of the Exchange on NetApp: 1. How can I create a "Exchange consistent" snapshots to provide for the backup of Exchange (snapshot) with of the shelf software like Legato Networker (no Exchange API - just files). 2. As circular logging should be disabled - Is there any API/procedure that will clean up the transaction logs as it is done by "Exchange aware" backup clients but without actually doing the backup to a physical device. 3. In case of a need for partial data recovery - is there a Filer local copy, the convert a part of a snapshot from RO to RW mode besides a through the host copy can can always be done but consumes a lot of bandwidth, time and resources. Would appreciate any thoughts/suggestions/enlightnings/pointers to the above subjects. -- Itzik Meirson <mailto:imeirson@xpert.com> Xpert Integrated Systems Systems & Storage Integration Specialist

1 0

Re: Dual power on F760
by sirbruce＠ix.netcom.com 13 May '99

13 May '99

On 05/11/99 10:39:39 you wrote: > >On Mon, 10 May 1999, Timothy Moore wrote: > >> The machine 'hitz' is a F760 in my lab running 5.3. At the back of the machine I >> turned one power supply off then back on, repeat with the other. The messages were >> generated to the console, system logger and as event notification email. > >The machines I was referring to, are NetCaches C760 running 3.3R2D12. We >did the same with switching the power supplies on and off and the Ok >message never came, only the failed one did. Sounds like a bug. Contact support. :) Bruce

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