Re: Fibre Channel Questions

Subject: Re: Fibre Channel Questions

From: David Kaplowitz <email@hidden>

Date: Thu, 1 Jun 2006 17:32:09 +0100

Delivered-to: email@hidden

thanks for the reply

try grepping system.log for Fusion, and you'll probably see some Reset lines

I get:

May 24 18:21:47 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

May 24 18:21:47 daughtership kernel[0]: FusionMPT: External Bus Reset for

SCSI Domain = 0

May 24 18:31:12 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

May 24 18:31:12 daughtership kernel[0]: FusionMPT: External Bus Reset for

SCSI Domain = 0

May 24 18:32:16 localhost kernel[0]: Resetting IOCatalogue.

May 24 18:39:14 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

May 24 18:39:14 daughtership kernel[0]: FusionMPT: External Bus Reset for

SCSI Domain = 0

May 24 18:50:55 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

May 24 18:50:55 daughtership kernel[0]: FusionMPT: External Bus Reset for

SCSI Domain = 0

May 24 19:22:17 daughtership kernel[0]: jnl: journal start/end pointers

reset! (jnl 0x2c6ad7c; s 0x1a10000 e 0x1a10000)

May 24 19:32:50 daughtership kernel[0]: jnl: journal start/end pointers

reset! (jnl 0x2c6ac98; s 0x1a10000 e 0x1a10000)

May 24 19:33:45 daughtership kernel[0]: jnl: journal start/end pointers

reset! (jnl 0x2c6ad7c; s 0x371e00 e 0x371e00)

May 24 19:34:30 daughtership kernel[0]: jnl: journal start/end pointers

reset! (jnl 0x2c6a494; s 0x371e00 e 0x371e00)

May 24 19:58:35 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

May 24 19:58:35 daughtership kernel[0]: FusionMPT: External Bus Reset for

SCSI Domain = 0

May 24 20:00:42 daughtership kernel[0]: FusionMPT: Notification = 5

(External Bus Reset) for SCSI Domain = 0

..and so on...

so you're right!

start testing/swapping suspect components one by one if you can.

since my post, i've removed the backup machine from the mix completely, and installed NetVault onto the xServe, and things seem to be going better.

What version of firmware, and have you started seeing this problem since you upgraded? I noticed a lot of oddness with our fabric after the 1.5 update, and we're using switch's of the same lineage, so just curious.

1.5, and come to think of it, the problems did start after the upgrade to 1.5, although we also introduced the tape lib and backup software to the fibre mix at the same time...

So, the library is one port, and the drives are one port? Just trying to understand better, because I'd think each drive has it's own fibre connection, and the library's is separate, for a total of three. Also, the library will be set to a different speed and possibly topology than your drives. Exabyte is like that anyway.

The Scalar 24 has 2 tape drives, each with its own fibre port. These are connected to ports 3&4 in the Emulex 355. Same with xServe RAID.

- a machine running our backup software, Bakbone's NetVault, 1

fiber port

What is this machine, and how does the XServe fit into this then?

This machine has been removed. The xServe handles our AFP & OD, and now also runs NetVault backups at night.

all fiber devices are connected to an Emulex 355 fiber switch

I'm pretty sure you can't zone on this switch anyway. IIRC this

switch is basically the old Vixel 355/335, and you can't zone those.

You can indeed zone on the Emulex, in its IP-based GUI, but when I tried a couple things after hours, they didn't work at all. I tried setting up 2 zones, one with xServe and xServe RAID, and one with xServe and Scalar 24, and when I did (slight scare), the RAID disappeared from the dektop.

So it's as it was at the moment, with no zoning and LUN masking on between the xServe and xServe RAID.

SO, we seem to be getting there, but there's one other issue that's cropped up since I removed the backup machine, or rather, an issue that's continued:

I'm getting some "Ordered Set Errors" in the switch on port 5, which is where the xServe is plugged in. I haven't yet been able to diagnose fully, and it may just be that it appears when I dismount a partition of the xServe RAID as I did yesterday to run disk warrior, but i don't even know what they are - do you?

thanks again for the detailed and relevant reply.

david

------------------------

david kaplowitz / somethin' else

020 7204 1996

07903 324 617

somethin' else

units 1-4

1a old nichol street

london e2 7hr

------------------------

david kaplowitz / somethin' else

020 7204 1996

07903 324 617

somethin' else

units 1-4

1a old nichol street

london e2 7hr

_______________________________________________ Do not post admin requests to the list. They will be ignored. Macos-x-server mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macos-x-server/email@hidden This email sent to email@hidden