toasters January 2003

toasters@lists.teaparty.net

69 participants
78 discussions

SV: statit figures normal?
by Öberg Mats 08 Jan '03

08 Jan '03

Hi, yes, your statit figures look normal for that kind of configuration. That's the way WAFL works, it writes across the entire volume, equal amonts of data to each disk. If you mix drivesizes in a volume you will get this kind of behaviour, the smaller disks get filled first so writes go only to the bigger disks. (in your case the limit is roughly 67% of the volume). ---- Mats -----Ursprungligt meddelande----- Från: Kemal KURBAN [mailto:kemal.kurban@mynet.com] Skickat: den 8 januari 2003 18:07 Till: toasters(a)mathworks.com Ämne: statit figures normal? Hi, we've two F740 on one of them there is a volume consisting of 14x18GB disks + 14x36GB disks and it is %70 full we're using it to store our mail files, after peak hour only an additional script is running to delete mailboxes which are not used no more. here is the statit and sysconfig -r output LOOK FOR THE UTILIZATION of 18GB disks !!! Hostname: filer ID: xxxxxxxx Memory: 512 MB NetApp Release 5.3.7R2: Sat Dec 9 03:15:25 PST 2000 <1> Start time: Wed Jan 8 18:22:41 EET 2003 CPU Statistics 116.814359 time (seconds) 100 % 49.343172 system time 42 % 6.705551 rupt time 6 % (594640 rupts x 11 usec/rupt) 42.637621 non-rupt system time 37 % 67.471186 idle time 58 % 28.718686 time in CP 25 % 100 % 1.971005 rupt time in CP 7 % (163591 rupts x 12 usec/rupt) 16.135991 non-rupt system time in CP 56 % 10.611689 idle time in CP 37 % Miscellaneous Statistics (per second) 6860.76 hard context switches 0.00 spurious disk interrupts 1316.39 NFS operations 485.18 network KB received 1855.76 network KB transmitted 3989.96 disk KB read 1603.31 disk KB written 176.69 NVRAM KB written 434.37 WAFL bufs given to clients 156.27 checksum cache hits ( 35%) 0.00 no checksum - odd length 0.14 no checksum - partial buffer WAFL Statistics (per second) 411.75 name cache hits ( 57%) 308.39 name cache misses ( 42%) 4072.84 inode cache hits ( 92%) 314.68 inode cache misses ( 7%) 0.00 blocks read 433.80 blocks read-ahead 0.00 chains read-ahead 318.80 blocks written 68.43 stripes written 0.00 blocks over-written 0.06 wafl_timer generated CP 1.13 snapshot generated CP 0.00 wafl_avail_bufs generated CP 0.37 dirty_blk_cnt generated CP 0.00 full NV-log generated CP 0.00 back-to-back CP 92548 buffers Network Interface Statistics (per second) interface e0 0.91 receive packets 0.01 transmit packets 128.51 receive bytes 2.89 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full interface e3a 0.00 receive packets 0.00 transmit packets 0.00 receive bytes 0.00 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full interface e3b 0.00 receive packets 0.00 transmit packets 0.00 receive bytes 0.00 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full interface e3c 1667.27 receive packets 2019.44 transmit packets 406211.03 receive bytes 1753135.45 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full interface e3d 529.94 receive packets 496.12 transmit packets 90504.47 receive bytes 147176.58 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full interface vh 0.00 receive packets 0.00 transmit packets 0.00 receive bytes 0.00 transmit bytes 0.00 receive errors 0.00 transmit errors 0.00 collisions 0.00 packets dropped--queue full Disk statistics (over 116 seconds) ut% is the utilization % of the disk; chain is avg number of 4K blocks/op; usecs is round trip time to disk in microseconds per 4K block. disk ut% xfers/s urds/s chain usecs cwrs/s chain usecs crds/s chain usecs 1.0 12 14.95 14.95 1.17 7423 0.00 0.00 0 0.00 0.00 0 1.1 12 15.56 15.56 1.16 7192 0.00 0.00 0 0.00 0.00 0 1.2 11 13.03 13.03 1.23 7104 0.00 0.00 0 0.00 0.00 0 1.3 0 0.31 0.15 1.00 9666 0.15 1.00 0 0.00 0.00 0 1.4 12 14.25 14.25 1.16 7415 0.00 0.00 0 0.00 0.00 0 1.5 13 16.37 16.37 1.13 7160 0.00 0.00 0 0.00 0.00 0 1.6 13 16.19 16.04 1.14 7398 0.15 1.00 0 0.00 0.00 0 1.8 10 13.51 13.51 1.24 6504 0.00 0.00 0 0.00 0.00 0 1.9 11 13.50 13.50 1.19 7054 0.00 0.00 0 0.00 0.00 0 1.10 0 0.00 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 1.11 10 12.31 12.31 1.13 7656 0.00 0.00 0 0.00 0.00 0 1.13 10 12.65 12.65 1.17 7174 0.00 0.00 0 0.00 0.00 0 1.12 11 13.06 13.06 1.18 7216 0.00 0.00 0 0.00 0.00 0 1.14 11 13.18 13.18 1.17 7344 0.00 0.00 0 0.00 0.00 0 1.16 48 59.56 36.52 1.15 8775 11.07 2.39 3695 11.96 1.71 2770 1.17 48 60.43 36.48 1.15 8719 11.56 2.49 3592 12.37 1.74 2937 1.18 49 61.07 37.36 1.14 8780 11.66 2.35 3660 12.05 1.72 2866 1.19 50 60.90 37.85 1.15 8889 11.01 2.39 3838 12.03 1.80 3095 1.20 0 0.00 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 1.21 13 29.61 0.00 0.00 0 13.56 5.12 1556 16.04 2.68 836 1.22 49 60.86 37.44 1.15 8811 11.33 2.48 3579 12.07 1.76 2526 1.24 39 60.49 36.58 1.16 6911 11.22 2.35 3088 12.68 1.64 1935 1.25 38 59.46 34.81 1.15 7004 11.86 2.41 2850 12.79 1.57 1879 1.26 38 60.55 35.93 1.16 6766 11.92 2.31 3083 12.69 1.63 2041 1.27 39 62.05 37.09 1.18 6627 12.29 2.33 2963 12.66 1.66 2054 1.28 38 59.59 35.42 1.14 7053 11.20 2.25 3216 12.96 1.64 1908 1.29 39 59.53 35.09 1.13 7161 11.94 2.37 3002 12.49 1.62 2088 1.30 38 60.45 35.15 1.17 6701 12.18 2.34 2984 13.11 1.63 2041 Interrupt Statistics (per second) 1007.62 Clock 0.18 others 1007.81 total filer1*> sysconfig -r Volume vol0 (root) RAID group 0 RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) --------- ----- ------------ ---- -------------- -------------- parity 1.3 1 0 3 FC:A 17000/34816000 17366/35566480 data 1.6 1 0 6 FC:A 17000/34816000 17366/35566480 data 1.5 1 0 5 FC:A 17000/34816000 17366/35566480 data 1.1 1 0 1 FC:A 17000/34816000 17366/35566480 data 1.4 1 0 4 FC:A 17000/34816000 17366/35566480 data 1.0 1 0 0 FC:A 17000/34816000 17366/35566480 data 1.14 1 1 6 FC:A 17000/34816000 17366/35566480 data 1.9 1 1 1 FC:A 17000/34816000 17366/35566480 data 1.13 1 1 5 FC:A 17000/34816000 17366/35566480 data 1.8 1 1 0 FC:A 17000/34816000 17366/35566480 data 1.11 1 1 3 FC:A 17000/34816000 17366/35566480 data 1.2 1 0 2 FC:A 17000/34816000 17366/35566480 data 1.12 1 1 4 FC:A 17000/34816000 17366/35566480 RAID group 1 RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) --------- ----- ------------ ---- -------------- -------------- parity 1.21 1 2 5 FC:A 34500/70656000 35239/72170880 data 1.22 1 2 6 FC:A 34500/70656000 35239/72170880 data 1.19 1 2 3 FC:A 34500/70656000 35239/72170880 data 1.18 1 2 2 FC:A 34500/70656000 35239/72170880 data 1.16 1 2 0 FC:A 34500/70656000 35239/72170880 data 1.17 1 2 1 FC:A 34500/70656000 35239/72170880 data 1.24 1 3 0 FC:A 34500/70656000 35003/71687368 data 1.25 1 3 1 FC:A 34500/70656000 35003/71687368 data 1.26 1 3 2 FC:A 34500/70656000 35003/71687368 data 1.27 1 3 3 FC:A 34500/70656000 35003/71687368 data 1.28 1 3 4 FC:A 34500/70656000 35003/71687368 data 1.29 1 3 5 FC:A 34500/70656000 35003/71687368 data 1.30 1 3 6 FC:A 34500/70656000 35003/71687368 Spare disks RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) --------- ----- ------------ ---- -------------- -------------- spare 1.10 1 1 2 FC:A 0 17366/35566480 spare 1.20 1 2 4 FC:A 0 34732/71132960 Kemal KURBAN

1 0

740 cluster OS question
by Drew O'Donnell 08 Jan '03

08 Jan '03

We have a 740 cluster which we will be redeploying to production. The current GA release for this model is 6.22, however we were considering using 6.3.1 because this is the version we are running on our other filers. Does any one have any comments or recommendation on what we are considering? Thanks Drew

2 1

RE: accessing local group
by Bui, Marcus 07 Jan '03

07 Jan '03

I called Netapp support and this are the recommendations. It did fix the problem. Thanks for all the input. 1. Remove the machine account for the Filer from the WindowsNT domain controller. 2. On the Filer, type: filer> cifs terminate 3. Issue these commands: filer> mv /etc/lclgroups.cfg /etc/lclgroups.old filer> mv /etc/filersid.cfg /etc/filersid.old filer> mv /etc/cifsconfig.cfg /etc/cifsconfig.old 4. Run cifs setup, and enter your apropriate information 5. Verify that you can join the WinNT domain, and that your users can now access and map the shares (if they cannot, please let me know, and please tell me what error messages you get) 6. If you dont want to re-create your shares manually from the Domain controller, Computer Manager, you can: cifs terminate mv /etc/cifsconfig.old /etc/cifsconfig.cfg cifs restart Cheers, Marcus -----Original Message----- From: Reinoud Reynders [mailto:reinoud.reynders@uz.kuleuven.ac.be] Sent: Thursday, January 02, 2003 4:18 PM To: Bui, Marcus; 'Fox, Adam'; toasters(a)mathworks.com Subject: Re: accessing local group An other possibility is that you run again 'cifs setup'. This do the job. You are than of course some minutes without CIFS. Best regards, Reinoud UZLEUVEN Belgium ----- Original Message ----- From: "Bui, Marcus" <Marcus_Bui(a)AIMFUNDS.COM> To: "'Fox, Adam'" <Adam.Fox(a)netapp.com>; <toasters(a)mathworks.com> Sent: Thursday, January 02, 2003 10:25 PM Subject: RE: accessing local group > I can mapped to the ETC folder, but unable to view/change the contents. Get > 'access denied'. > > -----Original Message----- > From: Fox, Adam [mailto:Adam.Fox@netapp.com] > Sent: Thursday, January 02, 2003 2:49 PM > To: 'Bui, Marcus' > Subject: RE: accessing local group > > > The key is to be in the Local Adminstrators group. When you > changed domains, perhaps this wasn't updated. Try taking a peek > at the filer etc\lclgroups.cfg on the filer and look at the SIDs > under the Administrators group. You can use the cifs lookup command > to decode the SIDs. If the Domain Admins Group isn't there, fix it > by changing the SIDs. BE VERY CAREFUL and have your old copy of this > file available either by copying it or having a known good copy in a > snapshot because you want to be careful and be able to fall back. > In order to make CIFS see changes in this file, you must terminate and > restart cifs (at least it's quicker than a reboot). > > Hope this helps. You may want to consider calling NTAP to walk you > through it, but that's how I would approach it. > > -- Adam Fox > NetApp Professional Services, NC > adamfox(a)netapp.com > > > > -----Original Message----- > From: Bui, Marcus [mailto:Marcus_Bui@AIMFUNDS.COM] > Sent: Thursday, January 02, 2003 2:39 PM > To: 'toasters(a)mathworks.com' > Subject: accessing local group > > > All, > > I have a test F85 (ONTAP 6.2.2) in an NT4 Domain but I am unable to access > the local group on the filer. The box was previously in W2K AD but was > moved to an NT4 domain. The filer is listed in the domain. I logged as > Domain Admin but still get 'access denied' when trying to access the filer's > local group. Did 'cifs testdc' and it is authenticating to the correct DC. > > Any suggestions is appreciated. > > thanks, > > Marcus Bui > >

1 0

LDAP auth ?
by Frank Bonnet 07 Jan '03

07 Jan '03

Hi Advice to netapp folks Monthes ago I asked if Ontap will support native LDAP (openldap) authentication ... I know active directory is supported BUT we do not want use M$ stuff here, we have a fully functionnal LDAP server running under UNIX ( Linux/Debian) and are satisfied with it. I repost my request as I have for now a brand new F87 that is not in production for now ( it will be in few weeks ) This machine would be a test platform if there is some beta ontap release that it supposed to support this kind of LDAP authentication. Thanks for any answer. -- Frank Bonnet

2 1

UK User group
by Clawson, Simon 07 Jan '03

07 Jan '03

Hi, We signed up to the UK user group late last year, and I was wondering if anyone has any idea when the next meeting will be? Simon Simon Clawson Mentor Graphics HDL Designer Series R&D Rivergate Systems Administrator London Road Berkshire RG14 2QB TEL: 01635 811409

1 0

Re: a what-if disaster recovery question?
by ScottUCE＠aol.com 07 Jan '03

07 Jan '03

Extra filer heads not only make great swap/failover boxes, but excellent on-site spares kits and test beds as well. As such, many of our customers are dropping NetApp's expensive hardware support on their F700's.........when they suffer a hardware failure, they can swap the failing part immediately; we then ship a replacement part that goes back into their failover filer/spares kit. F720, 740 and 760 filer heads (without licensing) can be purchased for pennies on the dollar :-) Scott Scott Fischmann Union Computer Exchange, Inc. 6233 Idylwood Lane Minneapolis, MN 55436 email: scott(a)unioncomputer.com phone: 952 935 7282 | fax: 952 935 5056 "Helping our customers make each dollar go further -- since 1991!" www.unioncomputer.com > Subj:Re: a what if disaster recovery question? > Date:1/6/2003 4:34:48 PM Central Standard Time > From:<A HREF="mailto:rb@telenor.net">rb(a)telenor.net</A> > To:<A HREF="mailto:DFinn@studentadvantage.com">DFinn(a)studentadvantage.com</A> > CC:<A HREF="mailto:toasters@mathworks.com">toasters(a)mathworks.com</A> > Sent from the Internet > > > > Daniel Finn <DFinn(a)studentadvantage.com> writes: > > > What if we have a failure in one of the heads of our F760s? We have a > spare > > F740 head, would we be able to just put that in place? Is the raid > > information stored on the disk or on firmware in the head? I'm mostly > > wondering what would happen to the data in this situation. > > All relevant raid and firmware information is stored on multiple > disks, so swapping heads is simple and straitforward. > > I have never swapped from a "larger" head to a "smaller" head, but > both heads are Alpha based, and it should work, atleast in theory, as > the HW is is more or less equivilant, but there are HW constraints on > max raw disk capasity, raw file system size, number of FC Host Bus > Adapters etc. These factors may trip an apparent easy swap - though. > > I have, on a couple of occations upgraded from F720 to F740 without > much ado. Only make sure to edit /etc/rc file ahead of shutdown. > > We keep a spare head for our F740s, but since the major load now is on > F8xx (soon to add a FAS960c), we won't keep a spare F740 head on location > anymore - not really important ... > > If disaster recovery is important, I would suggest a 'spare kit' or as > we do, service contract with a local agent with 24h delivery. > > > Also, are shelves swappable between F760s and F740s? > > Definitely yes, with or without disks > > >>>>>>>>>><.rune > > -- > Rune Bakken > Senior System Administrator > Telenor Business Solutions (Nextra), ManagedServices/*/IP-Services > Ph: +47 992 86 474 >

1 0

Re: a what if disaster recovery question?
by Rune Bakken 06 Jan '03

06 Jan '03

Daniel Finn <DFinn(a)studentadvantage.com> writes: > What if we have a failure in one of the heads of our F760s? We have a spare > F740 head, would we be able to just put that in place? Is the raid > information stored on the disk or on firmware in the head? I'm mostly > wondering what would happen to the data in this situation. All relevant raid and firmware information is stored on multiple disks, so swapping heads is simple and straitforward. I have never swapped from a "larger" head to a "smaller" head, but both heads are Alpha based, and it should work, atleast in theory, as the HW is is more or less equivilant, but there are HW constraints on max raw disk capasity, raw file system size, number of FC Host Bus Adapters etc. These factors may trip an apparent easy swap - though. I have, on a couple of occations upgraded from F720 to F740 without much ado. Only make sure to edit /etc/rc file ahead of shutdown. We keep a spare head for our F740s, but since the major load now is on F8xx (soon to add a FAS960c), we won't keep a spare F740 head on location anymore - not really important ... If disaster recovery is important, I would suggest a 'spare kit' or as we do, service contract with a local agent with 24h delivery. > Also, are shelves swappable between F760s and F740s? Definitely yes, with or without disks >>>>>>>>>><.rune -- Rune Bakken Senior System Administrator Telenor Business Solutions (Nextra), ManagedServices/*/IP-Services Ph: +47 992 86 474

1 0

a what if disaster recovery question?
by Daniel Finn 06 Jan '03

06 Jan '03

What if we have a failure in one of the heads of our F760s? We have a spare F740 head, would we be able to just put that in place? Is the raid information stored on the disk or on firmware in the head? I'm mostly wondering what would happen to the data in this situation. Also, are shelves swappable between F760s and F740s? Thanks Dan

2 1

Odd log entry
by Stephen C. Woods 06 Jan '03

06 Jan '03

We're stumped, it appears as though someone issued a disk swap command on one of our filers, however none of us did. NetApp Release 5.1.2R3 on a 540. Is there some other reason that this might occur, disk or SCSI bus glitch? As far as we can tell nothing has changed in the disk farm (syconfig -r matches the weekly email from this week). <scw> Thu Jan 2 10:08:52 PST [disk_config_admin]: *** NOTICE *** A disk has been swapped (removed or added) to a modular storage shelf. The system will wait 15 seconds and then check the status of all disk drives. Thu Jan 2 10:09:26 PST [disk_config_admin]: *** NOTICE *** Disk unit status check has completed. ----- Stephen C. Woods; UCLA SEASnet; 2567 Boelter hall; LA CA 90095; (310)-825-8614 Unless otherwise noted these statements are my own, Not those of the University of California. Internet mail:scw@seas.ucla.edu

3 2

RE: mounting foreign root volume (multiple root volumes) causes p anick
by Chuck Tomasi 06 Jan '03

06 Jan '03

Vol status will tell you which one is the "root" (default boot) volume. To change it use vol options. You can have as many "stand by" root volumes as you like, but it will only boot from the marked "root" under the options column output from vol status. --Chuck -----Original Message----- From: steve(a)hq.newdream.net [mailto:steve@hq.newdream.net] Sent: Friday, January 03, 2003 12:52 PM To: toasters(a)mathworks.com Subject: mounting foreign root volume (multiple root volumes) causes panick Does anyone know if you can mount two (or more) root volumes on one filer and how tell it which on to boot from? I had a filer/shelf with a corrupt vol0. However there was user data on there I wanted to get at. So I mounted it on another test filer whose root volume was named boot, but the filer now complained about having two root volumes (boot and vol0). Thu Jan 2 21:09:00 GMT [fmmbx_instanceWorke:info]: Disk 4.1 is a primary mailbox disk Thu Jan 2 21:09:00 GMT [fmmbx_instanceWorke:info]: Disk 4.0 is a primary mailbox disk Thu Jan 2 21:09:00 GMT [fmmbx_instanceWorke:info]: normal mailbox instance on primary side Thu Jan 2 21:09:01 GMT [raid.assim.tree.multipleRootVols:error]: Volumes vol0 and boot both claim to be the root volume. 2 root volumes found: vol0 boot PANIC: 2 root volumes found, 2 of which are online. in process rc on release NetApp Release 6.3.1 on Thu Jan 2 21:09:02 2003 version: NetApp Release 6.3.1: Wed Nov 20 13:00:54 PST 2002 cc flags: 1 I was trying to boot both as readonly so I could boot off a previous inode on the corrupt vol0. Anyone know if there is a way to specify which volume to boot from when you are mounting multiple root volumes on one filer? I've already gotten the data in another fashion but I thought it should be able to mount two root volumes on one filer. Steve

1 0

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 January 2003