Mailing List Archive

On 2002-09-12 (11:21, -0400), Dan O'Brien <dmobrien@lcsi.net> wrote:

> We recently became caretaker of a new to us F630. This unit has 6 disk
> shelves, each shelf loaded with 9GB drives. We had NetApps come in and
> set it up (recertify it) after it was shipped to us from GB (we are in
> Ohio). Everything went fine except we had a couple of bad power
> supplies and a bad SCSI cable to one of the shelves.
>
> We are now looking at maintaining the hardware ourselves. We are
> planning on securing some spare parts and build a small hot spare system
> from which we can pull parts as needed.
>
> Anyone have any suggestions as new owners of a F630? We have lots of
> Unix and PC experience, but this is the first hands on with a NetApps.

Things that might help you:

- handbooks shipped with your filer
- filer man pages (for the unix guys located at http://filer/na_admin/)
- filer docs (located at http://filer/na_admin)
- now.netapp.com
- Network-Appliance trainings (ask your local key account manager)

I dont know what to suggest you - go and play with it before it gets a hot
system.

> (BTW: why is NetApps so secretive with info on their web site? Docs,
> manuals, specs, are all highly guarded with logins, passwords, etc.)

e.g.: a secured web site can't get spidered by search engines

However, netapp grants access to everyone who wants to with a guest
account - its not meant to keep you away from all the knowledge in there
;-)


Greets,

--
Stefan Funke | Arcor Online GmbH
Betriebs- und Systemadministration | Stresemannallee 30
stefan.funke@arcor-online.net | D-60596 Frankfurt / www.arcor.de
Phone#: +49 69 633 97-0 | fax: +49 69 633 97-444

Dan O'Brien wrote:

> We are now looking at maintaining the hardware ourselves. We are
> planning on securing some spare parts and build a small hot spare
> system from which we can pull parts as needed.

Hi, remember us? We secured enough spare parts to build a small hot
spare F630 system, or at least we think we have :-) but are having some
trouble getting the single disk shelf working.

Whenever we physically attach the disk rack to the controller via the
SCSI cable, we get a SCSI bus timeout during the "scanning for disk
drives" phase. This happens whether the disk rack is powered up or not
or disk drives (Seagate ST39173WC (9.1GB) 3 1/2" 7200 RPM Single Ended
Ultra-1 SCSI) plugged in or not

The controller says it's a dual-diff-wide scsi, and we have it in slot
7. There is a footnote in some of the documentation for the disk shelf
that says:

"The first Dual StorageShelf 1 or 2 disk shelf in an F630 shelf pair
must be configured with a differential to single-ended SCSI converter.
Converters for these disk shelves are not interchangeable."

Ok, we're neophytes. What are we doing wrong? Is there some part
missing? The disk shelf itself, looks like the second disk shelf of a
pair if we look inside our existing F630 filer.

Thanks in advance,

Dan O'Brien, dmobrien@lcsi.net
Cell: 614-783-4859 Work: 614-476-8473
Home: 740-927-2178 Pataskala, OH

On 2002-10-10 (16:44, -0400), Dan O'Brien <dmobrien@lcsi.net> wrote:

> Dan O'Brien wrote:
>
> > We are now looking at maintaining the hardware ourselves. We are
> > planning on securing some spare parts and build a small hot spare
> > system from which we can pull parts as needed.
>
> Hi, remember us? We secured enough spare parts to build a small hot
> spare F630 system, or at least we think we have :-) but are having some
> trouble getting the single disk shelf working.

We never had a F630 but I guess you're having a FC[7..9] shelf with 7 bays
for disks and 2 for power. To get one of those shelfs running with the
minimum of disks you need at least 2 disk sitting in bay 3 and 4. (someone
correct me if I'm wrong)



--
Stefan Funke | Arcor Online GmbH
Betriebs- und Systemadministration | Stresemannallee 30
stefan.funke@arcor-online.net | D-60596 Frankfurt / www.arcor.de
Phone#: +49 69 633 97-0 | fax: +49 69 633 97-444

The key is the converter frob to allow a DIFF card to drive SE disks.

Change card, change disks (and shelf) or get the converter frob.


--
-Mark ... an Englishman in London ...

Dan O'Brien wrote:
> Dan O'Brien wrote:
>
> > We are now looking at maintaining the hardware ourselves. We are
> > planning on securing some spare parts and build a small hot spare
> > system from which we can pull parts as needed.
>
> Hi, remember us? We secured enough spare parts to build a small hot
> spare F630 system, or at least we think we have :-) but are having some
> trouble getting the single disk shelf working.

Yup, us again. We got a "frob" device to allow the differential scsi
controlller to talk to the single ended disk shelf and we are past that
hurdle.

We are now running into a problem after we've zero'ed the disks, loaded
the software via NFS and "download" the boot blocks, and then on the
reboot we get this:

<typed from notes>

Starting NetApp Release 5.1.2...
Scanning....
Configuring disk drives: 7a.13 7a.0 7a.2 7a.15 7a.1 7a.12
<some log message header> Cluster Monitor: no interconnect found
<some log message header> WARNING: Cluster Monitor operating in degraded
mode.
<some log message header> Cluster Monitor: both partner mailbox disks
have failed
Disk 7a.0 is reserved for "hot spare"

1 disk is reserved for "hot spare"

Can't have SCSI disks in Clustered Failover configuration
Illegal configuration. Halting
ok

This is a single full shelf of 7 9GB scsi drives.

After zeroing the disks it asked me how many disks to add to the root
volume as there were already 2 assigned, plus it said there were 4
spare, so I let it add the recommended 3 to give a full RAID 5 disk
group (or something like that). 4+2 = 6 so I have one "missing"?

Anyway, anything obvious I can fix? There's really nothing in the old
manuals I have that talks about fresh, clean install from boot floppy.


Thanks in advance,

Dan O'Brien, dmobrien@lcsi.net
Cell: 614-783-4859 Work: 614-476-8473
Home: 740-927-2178 Pataskala, OH

On Tue, Oct 29, 2002 at 04:54:43PM -0500, Dan O'Brien wrote:
> Dan O'Brien wrote:
> >Dan O'Brien wrote:
> >
> > > We are now looking at maintaining the hardware ourselves. We are
> > > planning on securing some spare parts and build a small hot spare
> > > system from which we can pull parts as needed.
> >
> >Hi, remember us? We secured enough spare parts to build a small hot
> >spare F630 system, or at least we think we have :-) but are having some
> >trouble getting the single disk shelf working.
>
> Yup, us again. We got a "frob" device to allow the differential scsi
> controlller to talk to the single ended disk shelf and we are past that
> hurdle.
>
> We are now running into a problem after we've zero'ed the disks, loaded
> the software via NFS and "download" the boot blocks, and then on the
> reboot we get this:
>
> <typed from notes>
>
> Starting NetApp Release 5.1.2...
> Scanning....
> Configuring disk drives: 7a.13 7a.0 7a.2 7a.15 7a.1 7a.12
> <some log message header> Cluster Monitor: no interconnect found
> <some log message header> WARNING: Cluster Monitor operating in degraded
> mode.
> <some log message header> Cluster Monitor: both partner mailbox disks
> have failed
> Disk 7a.0 is reserved for "hot spare"
>
> 1 disk is reserved for "hot spare"
>
> Can't have SCSI disks in Clustered Failover configuration
> Illegal configuration. Halting
> ok
>
> This is a single full shelf of 7 9GB scsi drives.
>
> After zeroing the disks it asked me how many disks to add to the root
> volume as there were already 2 assigned, plus it said there were 4
> spare, so I let it add the recommended 3 to give a full RAID 5 disk
> group (or something like that). 4+2 = 6 so I have one "missing"?
>
> Anyway, anything obvious I can fix? There's really nothing in the old
> manuals I have that talks about fresh, clean install from boot floppy.

Is there a 'card' in the unit that has nothing plugged in that is not
the NVRAM card?

Perhaps a cluster card...

(yes, I have no bananas eh experience with this)

--
Mike Horwath Admin & Manager @ VISI.com WORK: drechsau@visi.com
IRC: Drechsau http://www.visi.com/ HOME: drechsau@geeks.org
The only Minnesota ISP with public statistics: http://noc.visi.com/
Garbage In -- Gospel Out. - berkeley fortune(6)

At 16:31, on Oct 29, 2002, Mike Horwath wrote:
> On Tue, Oct 29, 2002 at 04:54:43PM -0500, Dan O'Brien wrote:
...
> > <some log message header> Cluster Monitor: both partner mailbox disks
> > have failed
> > Disk 7a.0 is reserved for "hot spare"
> >
> > 1 disk is reserved for "hot spare"
> >
> > Can't have SCSI disks in Clustered Failover configuration
> > Illegal configuration. Halting
> > ok
> >
> > This is a single full shelf of 7 9GB scsi drives.
> >
> > After zeroing the disks it asked me how many disks to add to the root
> > volume as there were already 2 assigned, plus it said there were 4
> > spare, so I let it add the recommended 3 to give a full RAID 5 disk
> > group (or something like that). 4+2 = 6 so I have one "missing"?
> >
> > Anyway, anything obvious I can fix? There's really nothing in the old
> > manuals I have that talks about fresh, clean install from boot floppy.
>
> Is there a 'card' in the unit that has nothing plugged in that is not
> the NVRAM card?
>
> Perhaps a cluster card...

If what you're looking to do is make it forget that it's part of a
cluster, you can try the following (which worked for us on some 760c's -
I've never used/seen a 600 series though). At the ok prompt, do
unsetenv partner-sysid


Davin.

Davin Milun wrote:
> At 16:31, on Oct 29, 2002, Mike Horwath wrote:
>
>>On Tue, Oct 29, 2002 at 04:54:43PM -0500, Dan O'Brien wrote:

>>><some log message header> Cluster Monitor: both partner mailbox disks
>>>have failed
>>>Disk 7a.0 is reserved for "hot spare"
>>>
>>>1 disk is reserved for "hot spare"
>>>
>>>Can't have SCSI disks in Clustered Failover configuration
>>>Illegal configuration. Halting
>>>ok

> If what you're looking to do is make it forget that it's part of a
> cluster, you can try the following (which worked for us on some 760c's -
> I've never used/seen a 600 series though). At the ok prompt, do
> unsetenv partner-sysid

BINGO!!! Hold your cards!!!

This bit of magic cleared the "cluster" FUBAR!

Many thanks for this specific suggestion and for all the help provided
by others.

Cheers,

Dan O'Brien, dmobrien@lcsi.net
Cell: 614-783-4859 Work: 614-476-8473
Home: 740-927-2178 Pataskala, OH