toasters

toasters@lists.teaparty.net

13515 discussions

options cift.home_dir question
by Jones, Bill 20 Nov '99

20 Nov '99

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 How do I turn it off??? It seemed like a good idea at the time, but now users are concerned that their homedirs are a share for everyone to see. The question load is getting very repetitive and annoying!!!! Thanks in advance. Bill -----BEGIN PGP SIGNATURE----- Version: PGP Personal Privacy 6.5.1 iQA/AwUBODX4t/poFdhjwhvjEQL8pgCfZx9qsvekgCGtS11y79pLADYCvRMAnijQ V88oVOycQszl87SJ2zFwC7F5 =kZjI -----END PGP SIGNATURE-----

1 0

Mailing list
by dennism＠dtn.com 19 Nov '99

19 Nov '99

Can I be added to the toasters mailing list? Thank you, dennis

1 0

SCSI bus timeout
by Dave Toal 19 Nov '99

19 Nov '99

Folks, Ignorance question here, if you don't mind. We're seeing this timeout error -- "relog syslog SCSI drive 6.0 max timeout count exceeded" followed by "Resetting SCSI bus 6 (ha #3)". However, sysconfig -r doesn't show drive 6.0 on ha id 3 anywhere. Could someone please explain what's going on here? Thanks much. Dave Stephen Lasseter wrote: > Fri Nov 19 00:00:01 GMT [smith: statd]: 12:00am up 25 days, 2:48 169239880 NFS ops, 0 CIFS ops, 63 HTTP ops > Fri Nov 19 01:00:01 GMT [smith: statd]: 1:00am up 25 days, 3:48 169345577 NFS ops, 0 CIFS ops, 63 HTTP ops > Fri Nov 19 02:00:01 GMT [smith: statd]: 2:00am up 25 days, 4:48 169451567 NFS ops, 0 CIFS ops, 63 HTTP ops > Fri Nov 19 02:20:08 GMT [smith: isp_timeout]: 6.0(0xfffffc0000e84180): command timeout, retrying request > Fri Nov 19 02:20:08 GMT [smith: isp_timeout]: Resetting SCSI bus 6 (ha #3) > Fri Nov 19 02:25:34 GMT [smith: isp_timeout]: 6.0(0xfffffc0000e84660): command timeout, retrying request > Fri Nov 19 02:25:34 GMT [smith: isp_timeout]: Resetting SCSI bus 6 (ha #3) > Fri Nov 19 02:32:05 GMT [smith: rc]: de_main: e0 : Link up. > Fri Nov 19 02:32:05 GMT [smith: rc]: relog syslog Fri Nov 19 02:30:53 GMT [smith: isp_timeout]: 6.0(0xfffffc0000e84b40): command timeout, retrying request > > Fri Nov 19 02:32:05 GMT [smith: rc]: relog syslog SCSI drive 6.0 max timeout count exceeded - system resetting to recover! > > Fri Nov 19 02:32:05 GMT [smith: rc]: NetApp Release 5.2.1P2 boot complete. Last disk update written at Fri Nov 19 02:30:22 GMT 1999 > > smith> sysconfig -r > Volume brs1 > > RAID group 0 > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > parity 0a.8 0a 1 0 FC:A 17000/34816000 17366/35566480 > data 0a.1 0a 0 1 FC:A 17000/34816000 17366/35566480 > data 0a.4 0a 0 4 FC:A 17000/34816000 17366/35566480 > data 0a.0 0a 0 0 FC:A 17000/34816000 17366/35566480 > data 0a.5 0a 0 5 FC:A 17000/34816000 17366/35566480 > data 0a.13 0a 1 5 FC:A 17000/34816000 17366/35566480 > data 0a.10 0a 1 2 FC:A 17000/34816000 17366/35566480 > data 0a.3 0a 0 3 FC:A 17000/34816000 17366/35566480 > data 0a.11 0a 1 3 FC:A 17000/34816000 17366/35566480 > data 0a.9 0a 1 1 FC:A 17000/34816000 17366/35566480 > data 0a.2 0a 0 2 FC:A 17000/34816000 17366/35566480 > data 0a.28 0a 3 4 FC:A 17000/34816000 17366/35566480 > data 0a.29 0a 3 5 FC:A 17000/34816000 17366/35566480 > data 0a.30 0a 3 6 FC:A 17000/34816000 17366/35566480 > > Volume spool (root) > > RAID group 0 > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > parity 0a.12 0a 1 4 FC:A 17000/34816000 17366/35566480 > data 0a.14 0a 1 6 FC:A 17000/34816000 17366/35566480 > data 0a.6 0a 0 6 FC:A 17000/34816000 17366/35566480 > > Volume inbox > > RAID group 0 > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > parity 0a.17 0a 2 1 FC:A 17000/34816000 17366/35566480 > data 0a.16 0a 2 0 FC:A 17000/34816000 17366/35566480 > data 0a.18 0a 2 2 FC:A 17000/34816000 17366/35566480 > data 0a.21 0a 2 5 FC:A 17000/34816000 17366/35566480 > data 0a.22 0a 2 6 FC:A 17000/34816000 17366/35566480 > data 0a.24 0a 3 0 FC:A 17000/34816000 17366/35566480 > data 0a.25 0a 3 1 FC:A 17000/34816000 17366/35566480 > > Volume hostapp > > 0a 1 4 FC:A 17000/34816000 17366/35566480 > data 0a.14 0a 1 6 FC:A 17000/34816000 17366/35566480 > data 0a.6 0a 0 6 FC:A 17000/34816000 17366/35566480 > > Volume inbox > > RAID group 0 > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > parity 0a.17 0a 2 1 FC:A 17000/34816000 17366/35566480 > data 0a.16 0a 2 0 FC:A 17000/34816000 17366/35566480 > data 0a.18 0a 2 2 FC:A 17000/34816000 17366/35566480 > data 0a.21 0a 2 5 FC:A 17000/34816000 17366/35566480 > data 0a.22 0a 2 6 FC:A 17000/34816000 17366/35566480 > data 0a.24 0a 3 0 FC:A 17000/34816000 17366/35566480 > data 0a.25 0a 3 1 FC:A 17000/34816000 17366/35566480 > > Volume hostapp > > RAID group 0 > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > parity 0a.20 0a 2 4 FC:A 17000/34816000 17366/35566480 > data 0a.19 0a 2 3 FC:A 17000/34816000 17366/35566480 > > Spare disks > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > spare 0a.27 0a 3 3 FC:A 0 17366/35566480 > spare 0a.26 0a 3 2 FC:A 0 17366/35566480 > > Partner disks > > RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys (MB/blks) > --------- ----- ------------ ---- -------------- -------------- > partner 1.0 1 0 0 FC:B 0 17366/35566480 > partner 1.1 1 0 1 FC:B 0 17366/35566480 > partner 1.2 1 0 2 FC:B 0 17366/35566480 > partner 1.3 1 0 3 FC:B 0 17366/35566480 > partner 1.4 1 0 4 FC:B 0 17366/35566480 > partner 1.5 1 0 5 FC:B 0 17366/35566480 > partner 1.6 1 0 6 FC:B 0 17366/35566480 > partner 1.8 1 1 0 FC:B 0 17366/35566480 > partner 1.9 1 1 1 FC:B 0 17366/35566480 > partner 1.10 1 1 2 FC:B 0 17366/35566480 > partner 1.11 1 1 3 FC:B 0 17366/35566480 > partner 1.12 1 1 4 FC:B 0 17366/35566480 > partner 1.13 1 1 5 FC:B 0 17366/35566480 > partner 1.18 1 2 2 FC:B 0 17366/35566480 > partner 1.19 1 2 3 FC:B 0 17366/35566480 > partner 1.20 1 2 4 FC:B 0 17366/35566480 > partner 1.14 1 1 6 FC:B 0 17366/35566480 > partner 1.16 1 2 0 FC:B 0 17366/35566480 > partner 1.17 1 2 1 FC:B 0 17366/35566480 > partner 1.22 1 2 6 FC:B 0 17366/35566480 > partner 1.21 1 2 5 FC:B 0 17366/35566480 > partner 1.32 1 4 0 FC:B 0 17366/35566480 > partner 1.33 1 4 1 FC:B 0 17366/35566480 > partner 1.34 1 4 2 FC:B 0 17366/35566480 > partner 1.35 1 4 3 FC:B 0 17366/35566480 > partner 1.36 1 4 4 FC:B 0 17366/35566480 > partner 1.37 1 4 5 FC:B 0 17366/35566480 > partner 1.38 1 4 6 FC:B 0 17366/35566480 > smith> > -Steve

3 2

64-bit protocol ops counters?
by Brian Tao 19 Nov '99

19 Nov '99

Easy question... did anything change since 5.2.1 to give us 64-bit ops counters? A couple of our F740's are busy enough to roll 2^32 NFS ops every two weeks, and I get bogus stuff like: # rsh nfs3 nfsstat -l 10.35.10.2 tor-adm1-nfs NFSOPS = 759822365 (4294967087%) 10.35.10.3 tor-dev1-nfs NFSOPS = 79004302 (4294967275%) 10.35.10.17 tor-vs1-nfs NFSOPS = 2388554097 (528%) 10.35.10.18 tor-vs2-nfs NFSOPS = 3999698242 (82%) 10.35.10.19 tor-vs3-nfs NFSOPS = 1563148406 (4294966864%) 10.35.10.20 tor-vs4-nfs NFSOPS = 576153104 (4294967138%) 10.35.10.21 tor-vs5-nfs NFSOPS = 2690811725 (444%) 10.35.10.26 tor-fp1-nfs NFSOPS = 466426683 (4294967168%) Tue Nov 16 23:00:01 EST [nfs3: statd]: 11:00pm up 116 days, 2:53 4281632236 NFS ops, 0 CIFS ops, 0 HTTP ops Wed Nov 17 00:00:02 EST [nfs3: statd]: 12:00am up 116 days, 3:53 4291222987 NFS ops, 0 CIFS ops, 0 HTTP ops Wed Nov 17 01:00:01 EST [nfs3: statd]: 1:00am up 116 days, 4:53 4730533 NFS ops, 0 CIFS ops, 0 HTTP ops Wed Nov 17 02:00:01 EST [nfs3: statd]: 2:00am up 116 days, 5:53 12718144 NFS ops, 0 CIFS ops, 0 HTTP ops -- Brian Tao (BT300, taob(a)risc.org) "Though this be madness, yet there is method in't"

1 0

Re: NetApp CEO
by Graydon Dodson 19 Nov '99

19 Nov '99

FYI. I know from personal experience that Dan Warmenhoven (Netapp's CEO) is a regular reader of this list..... Graydon Dodson grdodson(a)lexmark.com Lexmark International Inc.

1 0

where can I get F330 parts?
by Jay Orr 18 Nov '99

18 Nov '99

I was wondering with the EOL on the F330's, we had been thinking of adding more space, but to do so we'd need : 1) Another SCSI card. 2) More memory to handle it. 3) Another Tray. Anyone know where I can find this stuff? New or used... Anyone out there scrapping theirs? ----------- Jay Orr Systems Administrator Fujitsu Nexion Inc. St. Louis, MO

1 0

No more 4GB drives available??
by Dane Jasper 17 Nov '99

17 Nov '99

I've got an F230 that I need disks for, but Netapp is telling me that they don't have any to sell. I need the space, and would like to buy disks. Does anyone have a source for these, or some to sell? I'm looking for like four to six drives. -- Dane Jasper Sonic.net, Inc. (707)522-1000 mailto:dane@sonic.net http://www.sonic.net/ Key fingerprint = A5 D6 6E 16 D8 81 BA E9 CB BD A9 77 B3 AF 45 53

3 2

case #79131, Very Poor Backup performances: 1Gb/H with Veritas Netbackup through NFS
by Pierre Maubert 16 Nov '99

16 Nov '99

Hi, We have several F330 in TIF, 2 of them have an enourmous number of small files: over 3.5M for one single volume. We have increase the maxfiles value to 4328504. We have currently a significative degradation of backup performances on these 2 servers: backup are done at 1.5Gb, on both server, wheras we reach 5Gb/h for the 3 others. Servers are all running Dataontap5.2.3, and are all identical. F330, 128Mb Ram, 24*4Gb disks, I wonder what kind of fine tuning can be done to speed up the backup. -- Regards, Pierre . +------------------------------------------------------+ | Pierre Maubert, Sys_Admin Texas Instruments France | | email:p-maubert@ti.com Tel: +33 49322.2681 | | Fax: +33 49202.4669 Pager: 0606498982 | +------------------------------------------------------+

11 22

RE: FC-AL: Onboard vs. PCI Adapter Card
by Norman, Paul 16 Nov '99

16 Nov '99

> -----Original Message----- > From: Shaun T. Erickson [mailto:ste@research.bell-labs.com] > Sent: Monday, November 15, 1999 8:27 PM > To: Norman, Paul > Cc: toasters(a)mathworks.com > Subject: Re: FC-AL: Onboard vs. PCI Adapter Card > > > "Norman, Paul" wrote: > > > > Over the last couple days, there has been some discussion > about on-board FC-AL > > in our filers. This summer, NetApp started shipping > filers with an FC-AL > > adapter card rather than the on-board FC-AL to improve > reliability and > > performance. We did this because we found errors which, > though recoverable, > > potentially effect performance and loop stability. The > problem effected a very > > small fraction, just a few percent, of the installed base. > > Thank you for your reply. However, you didn't address the main > concern/perception of many folks here: that NetApp apparently shipped > many F700 series systems with onboard FC-AL controllers that had hard > failures shortly after being put into production, after which NetApp Perhaps I missed some e-mails, but I didn't see anyone suggesting that there were "many F700" filers that had "hard failures shortly after being put into production". As I said the number of filers with instability problems was quite small. We've shipped on the order of 5,000 F700 series filers, the vast majority of which run fine on on-board FC-AL. About one percent of these exhibited instability on the on-board FC-AL and not all of those the instant they were put into production. All that being said, as I look back on how we handled this problem I think NetApp should have been more aggressive in alerting the installed base. You can't be a chicken little crying the sky is falling every time you have a hiccup in the production line, but you also have to recognize when an issue is not just temporary problem. > apparently started sending out new filers with the PCI card > pre-installed, to be used in place of the onboard controller. That's > quite different from what you just described. > > Can you elaborate on and clarify this issue for us, please? I think I've been pretty clear, but please feel free to call me tomorrow if you want to discuss further. 408 822 6288 is my direct line. > > TIA, > -ste >

1 0

dump to Solaris tapehost w/ATL L200 fails
by Dennis Chandler 16 Nov '99

16 Nov '99

I am trying to use dump on a Network Appliance 230 (ontap 4.3.3) to dump to a remote tapehost. The remote tapehost is Solaris 2.5.1 with an ATL PowerStore L200 ( 5 tape loader). I realize that dump has reached the EOT. However, I'm having trouble getting dump to go to the next tape. If I list the /dev/nrst* files there is no /dev/nrst2. Here is the error message. Dennis Chandler > filer> dump 0ufb tapehost:/dev/nrst1 126 / > ... > DUMP: 66% done, finished in 2:23 > DUMP: Remote tape device write: expected 129024, got 0 > DUMP: write failed > DUMP: The ENTIRE dump is aborted. > DUMP: deleting "snapshot_for_dump.9" snapshot.

2 2

Cybernetics tape drive
by Frank Gilmurray 15 Nov '99

15 Nov '99

We just got a F720 filer and we looking into various backup solutions. We thinking about a Cybernetics 8050 with a 15 tape library. The 8050 is an AIT-2 drive with a 50GB native capacity. I'm soliciting comments (good and bad) about this tape system. If you have one of these tape systems, which backup program are you using with it? Thanks, Frank -- Frank Gilmurray Sr Systems Administrator SRI International tel: (650) 859-3246 333 Ravenswood Avenue fax: (650) 859-4812 Menlo Park, CA 94025 email: frank(a)erg.sri.com

1 0

PCI FC-AL
by Jay Soffian 15 Nov '99

15 Nov '99

So here's what I'm wondering. If NA suspects that there are problems with the onboard FC-AL in deployed filers, why hasn't NA issued a field notice to this effect? Why hasn't NA sent all of its support customers who have affected models the PCI FC-AL controller kits? It appears that NA is waiting till a customer has problems with the onboard FC-AL before shipping the PCI FC-AL kit. I know NA values its customers' data more than this. j. -- Jay Soffian <jay(a)cimedia.com> UNIX Systems Engineer Cox Interactive Media

2 1

failing SCSI disks
by Priebe, Jason 15 Nov '99

15 Nov '99

This is a really stupid question, but I can't find any answers to it in the documentation. We are going to replace the 4GB drives in our F210's shelf with 9GB drives by manually failing the 4GB disks one at a time. What concerns me is that the documentation mentions that failing a disk _permanently_ marks it as failed, so that it cannot be used in another filer. Am I totally naive, or is there some way that you can permanently mark a disk so that even if you reformat it, a filer still knows that the disk is failed? I don't intend to try to use the disks in another filer, but I am worried about rendering seven 4GB disks unusable with any SCSI system. Will I be able to use these disks with SCSI-equipped PCs? Jason Priebe WRAL OnLine http://www.wral-tv.com/

4 5

Considering filers for major deployment....
by Alan R. White 14 Nov '99

14 Nov '99

Hi All, I've got a toaster onsite ready to start evaluating. If everything works out OK there's terabytes of business for NetApp at our shop. I'm hoping the collective experience on the group can help us get up to speed with filers - the sales guys are really good about answering the questions but I suspect you're going to be a tad more impartial ;-) Hope you don't mind the newbie questions below - if you prefer mail me directly and I can summarise back to the list. The claims of simplicity, reliability, minimal downtime, performance for filing (compared to NT) and snapshots are what attracted me to look at the boxes - I haven't seen any horror stories in the archives - is this too good to be true? The FC-AL stuff recently looked a bit dodgey. How many folks actualy cluster their filers? Claims of 99.997% uptime without clustering sound, once again, too good. Is the clustering simple primary-failover or can we do n-way clusters with load sharing etc? Is the cluster setup really a one-command 15 minute job? User restore on their NT clients by mapping the snapshots looks a good idea. Is it usable in the real world? It would save us heaps of hassle with classic 'ask IT to do it' restores. Any good rule of thumb sizing advice for the amount of space to reserve for snapshots? Similarly for automated snapshot schemes, does anyone do multiple snapshots intra-day and maybe keep one for a longer period, e.g. keep a midnight snapshot for x days. Is SnapMirror up to the job of keeping an almost real-time remote replica, i.e. snap every minute if the networks up to it? Are there any operational issues around this stuff? Is anyone prepare to comment privately or otherwise on any recent comparisons they've done with Celera and Auspex? I understand the cost thing with EMC but loads of people seem to buy them still. This is not intended as flame bait for all the NetApp advocates. Any advice on what we should really include in our eval to really test the box out? ...or indeed comments in general that would be useful for us. TIA, Al

5 4

cf headache
by Jay Soffian 14 Nov '99

14 Nov '99

Well, adding the PCI FC-AL contollers that NetApp sent to replace the onboard FC-AL for our clustered F740 pair did _not_ go smoothly. This marks two times in a row now that I've attempted to use 'cf takeover' in order to prevent downtime and it has lead to downtime of over an hour. The previous time, we were bitten by the uptime bug. This time, who knows. When I tried to do a 'cf takeover' on filer A, for some reason, it decided to mark one of filer B's data disks as a hot-spare, and then it panic'ed: subzero> Sun Nov 14 02:56:35 EST [subzero: rc]: Cluster monitor: takeover initiated by operator Sun Nov 14 02:56:35 EST [subzero: cf_main]: Cluster monitor: UP --> TAKEOVER Sun Nov 14 02:56:35 EST [subzero: cf_takeover]: Cluster monitor: takeover started Sun Nov 14 02:56:48 EST [subzero: disk_admin]: Resetting all devices on ISP2100 in slot 1 Sun Nov 14 02:56:55 EST [subzero: cf_takeover]: Marking disk 1.8 as a "hot spare" disk.Sun Nov 14 02:56:56 EST [subzero: raid_disk_admin]: One disk is missing from volume partner:cim0a, RAID group 1. A "hot spare" disk is available and the missing disk will be reconstructed on the spare disk. PANIC: wafl_check_vbns: vbn too big on Sun Nov 14 07:56:57 1999 Cluster monitor: panic during takeover Cluster monitor: takeover capability will be disabled on reboot subzero came back up okay. I hoped the other filer (viking) would come back up okay and begin a RAID reconstuct, but it was convinced the filesystem was hosed: Disk 0a.4 is reserved for "hot spare" Disk 0a.8 is reserved for "hot spare" Disk 0a.13 is reserved for "hot spare" 3 disks are reserved for "hot spare". 27 disks are owned by clustered failover partner. Sun Nov 14 08:03:13 GMT [rc]: A disk is missing from one or more RAID groups. System starting in degraded mode. disk: RAID label 1 and 2: magic/time/gen/shutdown fsid/rgid/rgdn/total 0a.0 : RAID/942566217/1881/0 20036d6/0/3/17/ RAID/942566217/1881/0 20036d6/0/3/17/ 0a.1 : RAID/942566217/1881/0 20036d6/0/4/17/ RAID/942566217/1881/0 20036d6/0/4/17/ 0a.2 : RAID/942566217/1881/0 20036d6/0/1/17/ RAID/942566217/1881/0 20036d6/0/1/17/ 0a.3 : RAID/942566217/1881/0 20036d6/1/3/17/ RAID/942566217/1881/0 20036d6/1/3/17/ 0a.4 : SPARE/0/0/0 ffffffff/-1/-1/1/ SPARE/0/0/0 ffffffff/-1/-1/1/ 0a.5 : RAID/942566217/1881/0 20036d6/0/0/17/ RAID/942566217/1881/0 20036d6/0/0/17/ 0a.6 : RAID/942566217/1881/0 20036d6/0/5/17/ RAID/942566217/1881/0 20036d6/0/5/17/ 0a.8 : SPARE/0/0/0 ffffffff/-1/-1/1/ SPARE/0/0/0 ffffffff/-1/-1/1/ 0a.9 : RAID/942566217/1881/0 20036d6/1/5/17/ RAID/942566217/1881/0 20036d6/1/5/17/ 0a.10 : RAID/942566217/1881/0 20036d6/1/2/17/ RAID/942566217/1881/0 20036d6/1/2/17/ 0a.11 : RAID/942566217/1881/0 20036d6/2/4/17/ RAID/942566217/1881/0 20036d6/2/4/17/ 0a.12 : RAID/942566217/1881/0 20036d6/2/2/17/ RAID/942566217/1881/0 20036d6/2/2/17/ 0a.13 : SPARE/0/0/0 ffffffff/-1/-1/1/ SPARE/0/0/0 ffffffff/-1/-1/1/ 0a.14 : RAID/942566217/1881/0 20036d6/0/2/17/ RAID/942566217/1881/0 20036d6/0/2/17/ 0a.16 : RAID/942566217/1881/0 20036d6/2/3/17/ RAID/942566217/1881/0 20036d6/2/3/17/ 0a.17 : RAID/942566217/1881/0 20036d6/2/0/17/ RAID/942566217/1881/0 20036d6/2/0/17/ 0a.18 : RAID/942566217/1881/0 20036d6/2/1/17/ RAID/942566217/1881/0 20036d6/2/1/17/ 0a.19 : RAID/942566216/1659/0 36d6/0/0/2/P RAID/942566216/1659/0 36d6/0/0/2/P 0a.20 : RAID/942566216/1659/0 36d6/0/1/2/S RAID/942566216/1659/0 36d6/0/1/2/S 0a.21 : RAID/942566217/1881/0 20036d6/1/1/17/ RAID/942566217/1881/0 20036d6/1/1/17/ 0a.22 : RAID/942566217/1881/0 20036d6/1/0/17/ RAID/942566217/1881/0 20036d6/1/0/17/ 1.0 : SPARE/0/0/0 ffffffff/-1/-1/1/ SPARE/0/0/0 ffffffff/-1/-1/1/ 1.2 : RAID/942566451/3443/0 10022f5/0/4/6/ RAID/942566451/3443/0 10022f5/0/4/6/ 1.1 : RAID/942566451/3894/0 20022f5/0/1/18/ RAID/942566451/3894/0 20022f5/0/1/18/ 1.4 : RAID/942566451/3894/0 20022f5/2/4/18/ RAID/942566451/3894/0 20022f5/2/4/18/ 1.5 : RAID/942566451/3894/0 20022f5/2/1/18/ RAID/942566451/3894/0 20022f5/2/1/18/ 1.3 : RAID/942566451/3894/0 20022f5/2/5/18/ RAID/942566451/3894/0 20022f5/2/5/18/ 1.9 : RAID/942566451/3443/0 10022f5/0/0/6/ RAID/942566451/3443/0 10022f5/0/0/6/ 1.8 : RAID/942566451/3443/0 10022f5/0/5/6/ RAID/942566451/3443/0 10022f5/0/5/6/ 1.6 : RAID/942566451/3894/0 20022f5/0/3/18/ RAID/942566451/3894/0 20022f5/0/3/18/ 1.11 : RAID/942566451/3894/0 20022f5/1/5/18/ RAID/942566451/3894/0 20022f5/1/5/18/ 1.12 : RAID/942566451/3894/0 20022f5/2/0/18/ RAID/942566451/3894/0 20022f5/2/0/18/ 1.14 : RAID/942566451/3894/0 20022f5/1/4/18/ RAID/942566451/3894/0 20022f5/1/4/18/ 1.16 : RAID/942566451/3894/0 20022f5/2/2/18/ RAID/942566451/3894/0 20022f5/2/2/18/ 1.17 : RAID/942566451/3443/0 10022f5/0/2/6/ RAID/942566451/3443/0 10022f5/0/2/6/ 1.18 : RAID/942566451/3443/0 10022f5/0/3/6/ RAID/942566451/3443/0 10022f5/0/3/6/ 1.19 : RAID/942566451/3769/0 22f5/0/0/2/P RAID/942566451/3769/0 22f5/0/0/2/P 1.20 : RAID/942566451/3769/0 22f5/0/1/2/S RAID/942566451/3769/0 22f5/0/1/2/S 1.21 : RAID/942566451/3443/0 10022f5/0/1/6/ RAID/942566451/3443/0 10022f5/0/1/6/ 1.22 : RAID/942566451/3894/0 20022f5/0/4/18/ RAID/942566451/3894/0 20022f5/0/4/18/ 1.24 : RAID/942566451/3894/0 20022f5/0/0/18/ RAID/942566451/3894/0 20022f5/0/0/18/ 1.25 : RAID/942566451/3894/0 20022f5/0/5/18/ RAID/942566451/3894/0 20022f5/0/5/18/ 1.26 : RAID/942566451/3894/0 20022f5/0/2/18/ RAID/942566451/3894/0 20022f5/0/2/18/ 1.28 : RAID/942566451/3894/0 20022f5/1/2/18/ RAID/942566451/3894/0 20022f5/1/2/18/ 1.27 : RAID/942566451/3894/0 20022f5/1/1/18/ RAID/942566451/3894/0 20022f5/1/1/18/ 1.29 : RAID/942566451/3894/0 20022f5/1/0/18/ RAID/942566451/3894/0 20022f5/1/0/18/ 1.30 : RAID/942566451/3894/0 20022f5/1/3/18/ RAID/942566451/3894/0 20022f5/1/3/18/ 1.10 : RAID/942566451/3894/0 20022f5/2/3/18/ RAID/942566451/3894/0 20022f5/2/3/18/ Unclean shutdown in degraded mode without NVRAM protection. Filesystem may be scrambled. Please contact Network Appliance Customer Support. (You must boot from floppy to take any further action) Joy. I presume this is because subzero panic'd. Had subzero not panic'd I'm guessing subzero would have begun a reconstruct for viking's disks (after it blew-away that data disk for whatever reason). Anyway, at this point, we decided to put in the PCI FC-AL hoping the filer would notice the disks on reboot (I hadn't yet diagnosed why a data disk was missing). No go. So we ran wackz. 20 minutes later wackz found no problems. Rebooted the filer and it came up and began a reconstruct. It chose one of its other spares, not the former data-disk spare to reconstruct onto. (During reconstruct, the filer 'recovered error' twice on another disk in the same RAID group, glad it didn't decide to throw that disk or I guess I would have been trying to figure how to get that hot-spare-used-to-be-a-data-disk back as a data disk.) I didn't want to risk another 'cf takeover' (esp during a reconstuct), so I just halted the other filer to add its PCI FC-AL. No incident there. Case has been opened with Netapp. This is getting really old. Oh yeah, 5.2.3 on both filers. Both filers have succefully done a 'cf takeover' during the past (during testing a few different times). Apparently, 'cf takeover' only fails when you actually _need_ it. On the bright side, NetApp customer support was pretty helpful. It was a multi-national effort (I was coordinating from Atlanta, the person performing the maintenence at our co-lo was in Sunnyvale and the NA customer support person (Renee) was in the Netherlands). j.

1 0

FC-AL: Onboard vs. PCI Card - performance
by Jay Soffian 14 Nov '99

14 Nov '99

The replacement FC-AL PCI controller I'm about to add is a 32-bit card (as is the cluster FC-AL controller already installed). The on-board FC-AL controller is 64-bit. Are there any performance differences between the two? Geez, if this filer ever has to run in failover mode, it will have active FC-AL contollers in slots 1,2, active GigE in slots 6,7, active cluster interconnect in slot 8. Busy backplane. Also, suddenly, my F740s are down to one free slot. Hrmph. j.

1 0

Single-mode vif failover reaction time
by Brian Tao 13 Nov '99

13 Nov '99

I started playing with single-mode VIF's on 5.3.4 today (F740, QFE NIC installed in slot 3). I have e3a connected to one Cisco 2924XL and e3b connected to a different 2924XL. Both switches are bridged together so that it essentially looks like a 48-port switch to the hosts connected to it. adm-na5> vif create single na5-vif1 e3a e3b adm-na5> vif favor e3b adm-na5> ifconfig na5-vif1 10.35.8.21 netmask 255.255.255.128 up I created a single-mode trunk with e3a and e3b, favouring e3b (no particular reason, other than to see that the filer won't assume the first interface in the "vif" command will be favoured), ifconfigged it, and verified that it was pingable. I started up a few scripts to record any NFS outages while I played with the failover. Link failure was simulated by unplugging the filer's cable from both the switch end and the filer end. "vif status" and "vif stat" both verified that all traffic was passing over the e3b link exclusively. I yanked out the e3b cable from the switch: > Virtual interface (trunk) na5-vif1 > e3a e3b > Pkts In Pkts Out Pkts In Pkts Out > 294767 13473 856994 38899 > 0 0 3095 141 > 0 0 3236 149 > 0 0 2403 104 > Thu Nov 11 23:34:20 GMT [de2]: de_main: e3b : Link down. Check cable. > 2527 122 0 0 > 3737 162 0 0 > 3209 141 0 0 > 3551 156 0 0 Instantaneous failover! I was worried that the Catalysts would spend 20 or 30 seconds doing loop detection, but nothing even blinks at the loss of one of the links. Hurrah. :) But then I plug e3b back in, give it time to settle (at this point, the switch does do a recalc), and then unplug e3a to force a "give back" to e3b, I observe about 15 seconds where no traffic is passed over either link: > Virtual interface (trunk) na5-vif1 > e3a e3b > Pkts In Pkts Out Pkts In Pkts Out > 569739 25859 873099 39627 > 5558 239 0 0 > 3137 147 0 0 > 3550 162 0 0 > Thu Nov 11 23:35:40 GMT [de1]: de_main: e3a : Link down. Check cable. > 0 0 0 0 > 0 0 2 0 > 0 0 0 0 > 0 0 0 0 > 0 0 1 0 > 0 0 0 0 > 0 0 0 0 > 0 0 0 0 > 0 0 0 0 > 0 0 0 0 > 0 0 0 0 > 0 0 2 2 > 0 0 1 1 > 0 0 1 1 > 0 0 1 1 > 0 0 2347 103 > 0 0 1 1 > 0 0 1 1 > 0 0 1450 64 > 0 0 4857 206 > 0 0 3807 163 > 0 0 3983 174 NFS clients definitely notice that. Is this normal for the Netapp, or should I be looking at my switch configuration? Also, is there a way to force the Netapp to automatically switch back to the favoured link when it comes back up (perhaps after a short timeout, in case the interface is bouncing up and down)? -- Brian Tao (BT300, taob(a)risc.org) "Though this be madness, yet there is method in't"

2 2

FC-AL: Onboard vs. PCI Card
by Shaun T. Erickson 13 Nov '99

13 Nov '99

Last summer, our F540 failed, and we moved our data to an F740. Two days later, the system crashed everytime it hit 51% of a disk-rebuild. Ultimately, it was determined that the onboard FC-AL controller was bad and the system board was replaced with one from a slightly older F740. We've been fine ever since. NetApp sent us a PCI FC-AL card, "just in case", but it's still in the box, unused. Having followed the "AW: 'Loop break detected' followed by failover" thread, and noted that people seem to be using the PCI adapters now, instead of the onboard adapters, and that apparently NetApp is shipping them with the PCI adapter standard now, I'm in a quandry. Should I proactively install the PCI adapter and switch to it, or should I leave well enough alone? Can someone at NetApp comment on this and give me some guidance? TIA, -ste -- Desk: (908) 582-7629 Pager: (800) 756-1133 Cell: (908) 672-6456

3 2

AW: 'Loop break detected' followed by failover
by bentele＠mtu-friedrichshafen.com 13 Nov '99

13 Nov '99

Hi Jay, we also had a ' isp2100_timeout]: Resetting ISP2100 in slot ..' a few weeks ago. F740 , DOT 5.2.3, FC-AL 18 GB, Gbit Card. !! Be careful !! Replacing motherboard and FC-Adapter-Card at last solved the problem, after we replaced all cables, all shelves of this loop, and about 3 disks. Mit freundlichen Grüßen Markus Bentele MTU Friedrichshafen GmbH * 07541-90-2654 * bentele(a)mtu-friedrichshafen.com > ---------- > Von: Jay Soffian[SMTP:jay@cimedia.com] > Gesendet: Mittwoch, 10. November 1999 08:07 > An: toasters(a)mathworks.com > Betreff: 'Loop break detected' followed by failover > > > I've opened a case with NOW on this. > > One of my filers (F740/512MB/28x9GB disks/DOT 5.2.3) tonight just did > this: > > Tue Nov 9 19:00:01 EST [subzero: statd]: 7:00pm up 44 days, 13:45 > 2570782327 NFS ops, 0 CIFS ops, 0 HTTP ops > Tue Nov 9 19:16:46 EST [subzero: isp2100_main]: Loop break detected on > ISP2100 in slot 1. > Tue Nov 9 19:16:50 EST [subzero: isp2100_timeout]: Resetting ISP2100 in > slot 1 > Tue Nov 9 19:19:14 EST [subzero/viking: cf_takeover]: relog syslog Tue > Nov 9 19:17:15 EST [subzero: isp2100_timeout]: Resetting ISP2100 in slot > 1 > Tue Nov 9 19:19:14 EST [subzero/viking: cf_takeover]: relog syslog Tue > Nov 9 19:17:27 EST [subzero: isp2100_timeout]: Resetting ISP2100 in slot > 1 > Tue Nov 9 19:19:15 EST [subzero/viking: asup_main]: Cluster Notification > mail sent > > No core dump, I found the filer at an 'ok' prompt. > > I rebooted the filer, did a cf giveback a few minutes later and all > appears fine now. > > Anything recommended besides checking that all cables are tight and > all drives are fully seated? > > I couldn't find any bugs relating to this on NOW (except a bug in the > Diagnostics not resetting the FC-AL loop properly when a loop break > occurs). I've heard rumors that some F740's were shipped with bad > on-board FC-AL controllers, but this F740 has been in operation for > over a year w/o any problems (it's only been clustered since June > though). > > Suggestions? > > j. >

5 5

RE: failing SCSI disks
by Priebe, Jason 13 Nov '99

13 Nov '99

I _would_ dump, then replace all the disks, but we can't buy pre-canistered 9GB SCSI disks anymore. So we're (on the advice of our sales rep) replacing the disks in the canisters ourselves. This is a laborious process, and I don't want to add to the downtime between dump and restore by trying to do it offline, so we're trying to do it one by one while the filer is online. Are we crazy? Jason Priebe WRAL OnLine http://www.wral-tv.com/ -----Original Message----- From: Gordon Keegan To: Priebe, Jason; 'toasters(a)mathworks.com' Sent: 11/12/99 12:01 PM Subject: Re: failing SCSI disks > * Priebe, Jason (priebe(a)wral-tv.com) done spit this rhetoric: > > This is a really stupid question, but I can't find any > > answers to it in the documentation. We are going to replace > > the 4GB drives in our F210's shelf with 9GB drives by > > manually failing the 4GB disks one at a time. > > Not sure why you'd want to do that, as this would make only 4G of each of > the 9G drives useable... > What he said. When a smaller drive fails over to a larger spare, the larger spare is marked down as having the same capacity as the smaller. You would end up with a shelf of 9GB drives and the usable capacity of a shelf of 4GB drives. Better off dumping to tape and restoring to a shelf of empty 9GB's. -- Gordon Keegan, M10-10 Phone: (508) 261-4696 Motorola ING Fax: (508) 261-5757 20 Cabot Boulevard Mansfield, MA 02048 email: lgk011(a)dma.isg.mot.com

2 1

RE: Considering filers for major deployment....
by Sam Schorr 13 Nov '99

13 Nov '99

Alan, We have been using Snapmirror in production on 2 F740's and 2 F760's. We may be using it for different purpose than you might - its principally used to provide a RAID 1-ish protection for our on-line volumes, and a modified, limited read-only access for some of that data. The first thing to know is that Snapmirror uses Snapshot technology, so the first thing the filers do at the scheduled intervals is build a snapshot. It will always build a snapshot even if no data has changed, and it always copies a percentage of the volume being mirrored. This makes every-minute snapmirrors almost impractical for large volumes (ours are 200-250 gig). Are filers are "paired" using a directly connected gigE cross-connect (no switch or router). On a heavily loaded filer, snapmirrors can "timeout" as they try to build the snapshot - this is a known problem that netapp has a patch for that will be included in 5.3.4R1 (or so I'm told). We use a alternating, one hour window. That is, filer 1 in a pair runs at the 1 minute of the hour mark, while the other runs at the 30 minute of the hour mark. If you are snapping to a read-only filer, that has no other use (except for root), you get better results. Have fun! -----Original Message----- From: Mike Federwisch [mailto:mikef@netapp.com] Sent: Friday, November 12, 1999 3:44 PM To: Alan R. White Cc: toasters(a)mathworks.com Subject: Re: Considering filers for major deployment.... > Is SnapMirror up to the job of keeping an almost real-time remote replica, > i.e. snap every minute if the networks up to it? Are there any operational > issues around this stuff? This one I can answer. Keep in mind that SnapMirror is not targeted to be a "almost real-time" remote replica. Its performance depends greatly on network, load, filesystem size etc. I wouldn't expect it to be a real-time style replica. If the networks are up to it you can get 24M/sec over the wire. Snapshots are created for each transfer which also take time. So figure out how much data changes per minute how fast your networks are and how much load you are planning on and you start to see how fast the updates can happen. I have seen situations where a per minute schedule can be achieved but I would lean more towards every 15 minutes or so but really it depends on your situation. I think it is a great product (I am biased) but I want to make sure your expectations are set correctly. Mike Federwisch Network Appliance

1 0

RE: Considering filers for major deployment....
by Benn, Paul 13 Nov '99

13 Nov '99

Hi Al, There's a simplified description of SnapShots (with pretty pictures) in my NT backup paper: http://www.netapp.com/tech_library/3052.html A more detailed take on SnapShots as well as SnapRestore and SnapMirror can be found here: http://www.netapp.com/tech_library/3043.html As for comparing a filer to brands X and Y (assuming you can get eval units from X and Y), I would suggest doing some performance tests as well as manageability tests using standard NT tools. Set file permissions, create user groups on the filer, create shares etc.. If you install the SecureShare Access utility, try setting UNIX permissions from NT (really cool). If your intent is to use the multiprotocol (CIFS and NFS) functionality, I'd also recommend that you test file locking between CIFS and NFS clients. Regards, Paul Benn -----Original Message----- From: Alan R. White [mailto:arw@tipper.demon.co.uk] Sent: Friday, November 12, 1999 3:16 PM To: toasters(a)mathworks.com Subject: Considering filers for major deployment.... Hi All, I've got a toaster onsite ready to start evaluating. If everything works out OK there's terabytes of business for NetApp at our shop. I'm hoping the collective experience on the group can help us get up to speed with filers - the sales guys are really good about answering the questions but I suspect you're going to be a tad more impartial ;-) Hope you don't mind the newbie questions below - if you prefer mail me directly and I can summarise back to the list. The claims of simplicity, reliability, minimal downtime, performance for filing (compared to NT) and snapshots are what attracted me to look at the boxes - I haven't seen any horror stories in the archives - is this too good to be true? The FC-AL stuff recently looked a bit dodgey. How many folks actualy cluster their filers? Claims of 99.997% uptime without clustering sound, once again, too good. Is the clustering simple primary-failover or can we do n-way clusters with load sharing etc? Is the cluster setup really a one-command 15 minute job? User restore on their NT clients by mapping the snapshots looks a good idea. Is it usable in the real world? It would save us heaps of hassle with classic 'ask IT to do it' restores. Any good rule of thumb sizing advice for the amount of space to reserve for snapshots? Similarly for automated snapshot schemes, does anyone do multiple snapshots intra-day and maybe keep one for a longer period, e.g. keep a midnight snapshot for x days. Is SnapMirror up to the job of keeping an almost real-time remote replica, i.e. snap every minute if the networks up to it? Are there any operational issues around this stuff? Is anyone prepare to comment privately or otherwise on any recent comparisons they've done with Celera and Auspex? I understand the cost thing with EMC but loads of people seem to buy them still. This is not intended as flame bait for all the NetApp advocates. Any advice on what we should really include in our eval to really test the box out? ...or indeed comments in general that would be useful for us. TIA, Al

1 0

Re: failing SCSI disks
by Luke Mewburn 12 Nov '99

12 Nov '99

"Priebe, Jason" writes: > This is a really stupid question, but I can't find any > answers to it in the documentation. We are going to replace > the 4GB drives in our F210's shelf with 9GB drives by > manually failing the 4GB disks one at a time. > > What concerns me is that the documentation mentions that > failing a disk _permanently_ marks it as failed, so that it > cannot be used in another filer. Am I totally naive, or is > there some way that you can permanently mark a disk so that > even if you reformat it, a filer still knows that the disk > is failed? I did a similar thing a while ago to `migrate' disks in a volume to a different physical tray (I wanted to decommission a tray on an F330 to redeploy). To get the `failed' disks recognised by the filer again all I did was use the appropriate command on the boot floppies to zero the disk. I can't recall exactly, but I believe it's `disk_erase_label'. That command is available in the `rc_toggle_basic' mode of the filer. Make sure you use it on the right disk! NOTE: if you toast your data I ain't responsible! > I don't intend to try to use the disks in another filer, but > I am worried about rendering seven 4GB disks unusable with any > SCSI system. Will I be able to use these disks with > SCSI-equipped PCs? Yes, you should be able to. The `failed' state is just a flag in the label.

1 0

Re: Fast ENet cards for F330 and F540
by Luke Mewburn 12 Nov '99

12 Nov '99

"Walsh, Warren (MN65)" writes: > We've got two toasters, a F330 and F540. They're currently networked with > FDDI, but we would like to migrate to 100BaseT. NetApp has EOL'd these > models, even thought they're three years old, and does not sell these parts. > Anyone have suggestions for finding an alternate vendor for NetApp NICs? Or > a recommendation for a used/surplus dealer? How many 100Mb connections do you need per filer? The F330 has on-board 100Mb networking, although it's a different connector to the 10Mb network.

1 0

Two volumes with same name, DOT 5.2.1, how do I fix?
by John Stoffel 12 Nov '99

12 Nov '99

Hi guys, We're running DataOnTap 5.2.1 on a bunch of F330s. One of them had a disk failure this weekend, and since I was in for a power shutdown, I figured I'd just swap in a spare disk from a returned filer (upgrade to an F740). The problem is the before the old F330 was returned, we never did any of the disk commands to make all the disks into spares. We just copied the data off and shutdown the system. So I took one of these disks from the old Filer and did the usual disk swap, disk swap procedure to put it into the Filer with the failed disk. And now comes the problem. The Filer sees the new disk and scans it in, but hey! It's got volume info on it, and it won't initialize the disk and just use it on the new filer. Sigh... I've tried using the command 'disk fail 8a.4' but it won't let me because it's part of a volume. So I look at the volume status and I get this: swtoast> vol status Volume State Status Options vol0 online normal root vol0 offline foreign partial Yup, *two* volumes, both named 'vol0' and I can't get rid of the foreign one. If I just do 'vol status vol0' I get the online volume listed. What I want to do is just a volume destroy, or a disk initialize without any questions asked. But I can't figure out how to do this. Any hints? I do have a call into NetApp, but I've been on hold for about 15 minutes already... Thanks, John John Stoffel - Senior Unix Systems Administrator - Lucent Technologies stoffel(a)lucent.com - http://www.lucent.com - 978-952-7548 john.stoffel(a)ascend.com - http://www.ascend.com

4 6

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