toasters October 2001

toasters@lists.teaparty.net

136 participants
184 discussions

RE: Good experience with Netapp & exchange
by Jay Newton 19 Oct '01

19 Oct '01

I just finished migrating a 43 gig InfoStore to our 840 Filer. I created a new Exchange server and added it to the site. I then moved the new databases to the Filer and set up all the Snapshot stuff with SnapManager for Exchange. I then migrated all 679 mailboxes to the new server on the Filer and voila! 99% of the user population did not notice the difference. Address books and calendars pop up a little quicker as well. I then back up the Snapshot once a day to our tape library using CommVault Galaxy which has incredible Exchange support. There are a couple of really important things to remember when doing this. First, you really need a private network of 100 megabit or better for the Exchange server to talk to the Filer. You don't want to share traffic with everyone else. Use a crossover cable and use the IP in the path just to be 100% certain traffic is on the private network. Next, buy a copy of SnapManager for Exchange. It makes the whole thing work more smoothly. Finally, make sure you have another server attached to that private network to run Snap verifies. Snap verifies make sure the Snapshot you just took is consistent and restorable. You can Snap an Exchange database and defer verification until later and on another machine. That gives you about a 15 second window where performance marginally degrades every time you do a Snapshot. Our experience is 1 gig per minute doing a verify. Ours were taking 45 minutes and was slowing down the server pretty drastically. By moving to another server on the private network, we only incur the 15 second penalty instead of the 45 minute one. Not only does this run great, but test restores I have done to another server are incredibly fast. It takes longer to install the OS and Exchange than it does to restore the database! There was no data loss and logs rolled forward. -----Original Message----- From: netapp netappliance [mailto:netapp@mail.be] Sent: Friday, October 19, 2001 8:43 AM To: toasters(a)mathworks.com Subject: Good experience with Netapp & exchange I'm looking for good experience of migrating exchange to netapp ? ----------------------------------------------------- Mail.be, Free WebMail and Virtual Office http://www.mail.be

1 0

Good experience with Netapp & exchange
by netapp netappliance 19 Oct '01

19 Oct '01

I'm looking for good experience of migrating exchange to netapp ? ----------------------------------------------------- Mail.be, Free WebMail and Virtual Office http://www.mail.be

1 0

"unrecognized error 0xb" -- correct slot for tape scsi?
by Toal, Dave 19 Oct '01

19 Oct '01

Folks, I've been trying to get an F740 filer to recognize a dlt for several weeks. The error it returns is below. This same error also happens on another filer, also a 740. DLT checks out ok, according to the StorageTek tech's dignostics box. Have tried swapping cables, and using different dlt's -- _always_ this "unrecognized error". The filer has been rebooted many times during this process. mt fsf, bsf, weof, &c all work ok, according to mt status, but all tape writes fail. -- Dave Toal unix person T&T, nq, MA statler> dump u0f nrst0a /vol/spool DUMP: Dumping tape file 1 on nrst0a DUMP: creating "/vol/spool/../snapshot_for_backup.0" snapshot. DUMP: Using Full Volume Dump DUMP: Date of this level 0 dump: Fri Oct 19 02:02:52 2001. DUMP: Date of last level 0 dump: the epoch. DUMP: Dumping /vol/spool/ to nrst0a DUMP: mapping (Pass I)[regular files] DUMP: mapping (Pass II)[directories] Fri Oct 19 02:03:10 GMT [isp_main]: Tape (1.0.0):cmdblk 0xfffffc0000e020f8 op 0xa block 64512 aborted command (b 47, 0) DUMP: Tape write failed. Fri Oct 19 02:03:10 GMT [isp_main]: Tape (1.0) unrecognized error 0xbunrecognized error 0xb. DUMP: DUMP IS ABORTED DUMP: Deleting "/vol/spool/../snapshot_for_backup.0" snapshot. statler>

1 0

RE: which disk
by Mike Sphar 19 Oct '01

19 Oct '01

Well, at the very least you can just run in degraded mode while you get the correct disk. I would expect you could simply plug the other disk in. If it's the wrong one, the netapp should just ignore it, I would think. Then you just go about ordering a correct one. Or, the whole filer might go up in a big puff of smoke. Try it and let us know! (This is one reason why I prefer to always keep at least one spare available on all my filers. I prefer to let it grab a spare on its own, then I just have to show up at my convenience to swap in the replacement for the bad one.) -- Mike Sphar - Sr Systems Administrator - Engineering Support Peregrine Systems, Inc. -----Original Message----- From: Jim Harm [mailto:jharm@llnl.gov] Sent: Thursday, October 18, 2001 3:12 PM To: toasters(a)mathworks.com Subject: Re: which disk I did not make myself clear. the disks are in the shelves and I am going to fail one that is noisy with errors and remaps of sectors. I know which disk it is. I don't know if my spare, in my hand(SHZHB-ZZ), can replace the one in the shelf(ST15230N). I'd like to know before I fail the disk. Can I plug in a SHZHB-ZZ to replace a failed ST15230N? I can't event tell if a SHZHB-ZZ is a scsi or a 4 GB disk! We haven't had to replace a disk on this unit for over three years! All my production systems that I work on now are 760s and 840s. At 2:33 PM -0700 10/18/01, Chris Lamb wrote: > > >> How do I correlate the identification of the disk on the >> weekly report with the label on the disk? > > The report says 4 GB disk ST15230N and one of my disks says SHZHB-ZZ. >> Nowhere on the disk is there a designation of size, scsi or ultra, >> or even if it is scsi. > >Yow! That is old. :-) An F330, perhaps? What version of ONTAP are you >running? > >From the part number, that drive is a Seagate fast/narrow SCSI drive. >Have you looked at the front of the machine? Even on the old StorageWorks >SCSI shelves the red LED should light to indicate a failed drive... but >it's been a long, long time since I ran one of those. > >Good luck, > >-- Chris > -- }}}===============>> LLNL James E. Harm (Jim); jharm(a)llnl.gov System Administrator Compaq Clusters (925) 422-4018 Page: 423-7705x57152

2 1

Re: which disk
by Jim Harm 18 Oct '01

18 Oct '01

I did not make myself clear. the disks are in the shelves and I am going to fail one that is noisy with errors and remaps of sectors. I know which disk it is. I don't know if my spare, in my hand(SHZHB-ZZ), can replace the one in the shelf(ST15230N). I'd like to know before I fail the disk. Can I plug in a SHZHB-ZZ to replace a failed ST15230N? I can't event tell if a SHZHB-ZZ is a scsi or a 4 GB disk! We haven't had to replace a disk on this unit for over three years! All my production systems that I work on now are 760s and 840s. At 2:33 PM -0700 10/18/01, Chris Lamb wrote: > > >> How do I correlate the identification of the disk on the >> weekly report with the label on the disk? > > The report says 4 GB disk ST15230N and one of my disks says SHZHB-ZZ. >> Nowhere on the disk is there a designation of size, scsi or ultra, >> or even if it is scsi. > >Yow! That is old. :-) An F330, perhaps? What version of ONTAP are you >running? > >From the part number, that drive is a Seagate fast/narrow SCSI drive. >Have you looked at the front of the machine? Even on the old StorageWorks >SCSI shelves the red LED should light to indicate a failed drive... but >it's been a long, long time since I ran one of those. > >Good luck, > >-- Chris > -- }}}===============>> LLNL James E. Harm (Jim); jharm(a)llnl.gov System Administrator Compaq Clusters (925) 422-4018 Page: 423-7705x57152

1 0

which disk
by Jim Harm 18 Oct '01

18 Oct '01

I have a filer that is old and still used for testbeds. The disk failed and we have spares, but; How do I correlate the identification of the disk on the weekly report with the label on the disk? The report says 4 GB disk ST15230N and one of my disks says SHZHB-ZZ. Nowhere on the disk is there a designation of size, scsi or ultra, or even if it is scsi. It is from NetApp, but the only label from netapp on it is the test sticker. -- }}}===============>> LLNL James E. Harm (Jim); jharm(a)llnl.gov System Administrator Compaq Clusters (925) 422-4018 Page: 423-7705x57152

1 0

Re: ONTAP with RedHat linux
by Jered Heeschen 18 Oct '01

18 Oct '01

It's nice to hear I'm not the only one who's run into the hang with the 2.4 kernel, though it sounds like we get it more consistently. I've seen the same issue on a RedHat 6.2 box that's updated to the 2.4 kernel, which is why I'd point to something in the kernel rather than something RedHat-specfic. The best workaround seems to be using TCP to mount the NFS shares - that should allow for better performance than limiting readand write sizes to 1023. To do that, add "tcp" to the mount options, either in /etc/fstab or using the -o flag for "mount". You'd also need to turn on the nfs.tcp.enable option on the filer. Jered Steve Losen <scl(a)sasha.acc.virginia.edu>@mathworks.com on 10/18/2001 10:25:54 AM Sent by: owner-toasters(a)mathworks.com To: toasters(a)mathworks.com cc: Subject: Re: ONTAP with RedHat linux I am running RedHat 7.1 and NFS mounting a netapp F820 running 6.0.1R3. Things work pretty smoothly, but I have experienced the same problem as another poster (sorry, lost the message) where NFS writes occasionally delay and a "NFS server not responding" error is logged. I have not tried the rsize=1023,wsize=1023 work around. I have experienced the same problem with NFS servers other than Netapp, including a IBM RS6000 running AIX. Non Linux NFS clients are having no delay problems with the Netapp filers, so I'm pretty sure it is on the Linux end. For awhile I thought the NIC on my Linux box had botched the speed/duplex negotiation with the ethernet switch. Our network folks can tell me what speed the switch port has negotiated, but I don't know how to check my NIC from Linux. The network folks tell me they see no errors on my switch port when I experience NFS delays. I don't seem to have any other network problems, so I doubt this is the problem. I have been able to duplicate the problem on a second Linux box. However, on a third box, essentially identical to mine and also running RH 7.1, I could not duplicate it. A couple of things to be aware of with RedHat Linux and NFS. Prior to RedHat 7.0, only NFS v2 was supported. Furthermore, Linux only supported 16 bit uids and gids. We have a lot of files on our Netapp filers with uids greater than 65535. RedHat 7.1 supports NFS v3 and 32 bit uids and gids. 32 bit uids are critical in our environment. I administer our filers and when I was running RH 6.2, I could not get accurate file owner information with "ls" and I could not "chown" files to uids over 65535, so I had to use another unix box for filer administration. Now I can do it all from my Linux box. Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

RE: disk_fw_update
by Geoff Hardin 18 Oct '01

18 Oct '01

Thanks for the responses everyone! Jeff Sloan wrote: Yes there is definitely spin up issues with some of the old firmware on the 18G FCAL disks. Upgrading the firmware to the newest releases is the right move. Doing it individually is a very safe method up updating them, and definitely desirable, as you've had a few disk fail on you. I would like to note this is just on the older firmware revisions on the 18G drives. (hence once you finish upgrading you won't have to worry about this repeating itself). In regard to your current situation, The solution is to reboot the node that is reporting the mailbox disk error. We have a workaround that allows you to avoid rebooting, but it's only for ONTAP 6.1X and above. Did you halt the second system when upgrading the firmware? There's a specific process to upgrading the disk firmware on a cluster pair. You can review that process here: http://now.netapp.com/NOW/knowledge/docs/ontap/rel601r3/html/sag/fw_up5.htm… Cheers! :-) Jeff > -----Original Message----- > From: Geoff Hardin [mailto:geoff.hardin@dalsemi.com] > Sent: Thursday, October 18, 2001 1:40 PM > To: Sloan, Jeff > Subject: Re: disk_fw_update > > > Jeff, > The cluster is running DOT NetApp Release 5.3.4R2. We > are updating the firmware in anticipation of upgrading to > DOT 6.1.1R1. We have had problems with some of our filers > with the old 18GB disks during the upgrade; during the last > upgrade, we lost 3 disks, fortunately in 3 separate RAID > groups. Needless to say, it made for a very long, very > tense night as the disks rebuilt. We were hoping to update > the firmware of each disk separately so that we could avoid > this problem. > > Geoff Hardin > geoff.hardin(a)dalsemi.com > Politicians and diapers need changing often for similar reasons.

1 0

disk_fw_update
by Geoff Hardin 18 Oct '01

18 Oct '01

filer1> cf status filer1 has detected a mailbox disk error, takeover of filer2 disabled I have just finished upgrading the disk firmware on filer2. I had turned off clustering per the disk_fw_upgrade instructions prior to upgrading the disks. Once I had finished, I tried to turn clustering back on on filer1 and received the error message above. Any idea what's going on? filer2 does not report any errors on cf status. Thanks for the help! Geoff Hardin Dallas Semiconductor MIS UNIX geoff.hardin(a)dalsemi.com Politicians and diapers need changing often for similar reasons.

2 1

RE: quotas file
by Honeycutt, Michael 18 Oct '01

18 Oct '01

Thats how I do it also. For instance, if you want to use a scheduled task to do a quota off/on instead of resize, you could say: This would be first cmd file- rsh filername -l username:password -n quota off volumename call c:\foldername\filename.cmd This would be second cmd file- rsh filername -l username:password -n quota on volumename -----Original Message----- From: Jordan Share [mailto:iso9@phantasticant.com] Sent: Thursday, October 18, 2001 2:23 PM To: Grey, Michael; toasters(a)mathworks.com Subject: RE: quotas file Well, I can't help you with most of your questions, but I can tell you how I rsh from win2k to my filer. In my hosts.equiv file, I have a line like this: w.x.y.z Administrator Where w.x.y.z is the IP of the machine I am rsh'ing from. Then I use commands like this: rsh FILERNAME -l root:ROOTPASSWORD command_for_filer This works for me. Jordan -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com]On Behalf Of Grey, Michael Sent: Wednesday, October 17, 2001 8:12 PM To: 'toasters(a)mathworks.com' Subject: quotas file Well I have figured out how to set quotas based on global groups in Windows NT, but I'm having a little problem with the quotas file once completed. My script checks to see if a user is a member of a global group for a quota alocation (200MB, 350MB, and 500MB) via the "Global" command. This script runs automatically when a change is made to a quota global group. If the user ID is not in any global group then they would receive the default quota. The users who are members of a global group for quotas are written to a scratch file in the format of the quotas file. Once the script is complete the quotas file is deleted and then this scratch file is renamed to the quotas file. Everything works excellent but quota changes are not applied until I perform a "quota resize" command. I need to know if the filer automatically recalculates the quotas based on this file and if so, under what conditions? or, is there any way to have the filer recalculate quotas based on the quota file other then via the "quota resize" command. I'm trying to do this from a batch script under NT. I'm not sure how to RSH to the box via a script from NT if at all possible. Any help would be appreciated. Thanks.

1 0

← Newer
1
...
5
6
7
8
9
10
11
...
19
Older →

Jump to page:

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 October 2001