Exabyte 224 Magnum & Retrospect cleaning tape issue

[jpbuse]

I've got an Exabyte Magnum 224L FC library connected to my Mac xServe G5 running Retrospect (specs below). I have a cleaning tape in slot #1 of the library and I've assigned that slot as a cleaning slot within Retrospect. I *think* it is cleaning when it should. The main problem I have is when I view devices via Retrospect (Configure>Devices), slot #1 is indeed assigned as the cleaning slot, however the tape barcode in slot #1 according to Retro is a backup take, not the cleaning tape. The tape in slot #1 of the lib is the cleaning tape. What gives?

 

xServe G5 Dual 2Ghz, 6GB ram, OS X Server 10.4.9 universal

Retrospect 6.1.126, DAV:1.0.107, DUV:6.1.10.100

Connected via fibre channel through a QLogic 5602 FC switch

 

Cleaning slot turned on:

 

Cleaning slot turn off (showing the correct tape barcode):

[rhwalker]

I'll ask the two stupid questions first: Why aren't you using the industry-standard barcode for cleaning tapes of CLNxxxx ? Retrospect is happy with that for me. I think there is something magic about that designation. Did you ever use that barcode for a backup set, such that Retrospect is looking up that barcode and finding a match if it gets past the forced designation of that barcode in a "cleaning slot" when you turn off the cleaning slot designation?

 

Otherwise, two possibilities that I can think of. You aren't going to like the first one, because I've been there, done that, and it's painful.

 

(1) You may have a corrupt barcode database. Sadly, there is no way to "repair" or to "reload" or "edit" the barcode database. You can only clear it out and then have Retrospect relearn which tapes are associated with which barcodes by, one by one, initiating a "verify" on every barcoded tape you have. You don't have to verify the whole tape, but you do have to let Retrospect get going and read a couple of files at the start and then stop the backup. Took me about a week of loading and unloading tapes to get the barcodes re-learned. I turned in a feature request to either use an editable file for the database, or else to provide a screen or a utility for adjusting/entering the barcode data. I don't expect that feature request to ever be satisfied; there are more urgent "show stopper" issues that need to be addressed first.

 

Here's the link to the KB article:

Barcode database fixup KB article

 

(2) This might be related to a cosmetic bug with how Retrospect deals with empty library slots, displaying that there is a tape in an empty slot whose name is a copy of a tape in another slot. It's not something caused by wrong info returned by the library, because we have the Ethernet admin version of the Exabyte VXA-2 1x10 1u PacketLoader, so I can see what status flags, barcodes, etc. the autoloader thinks it has. Here is the detail of that bug, which I also turned in under our support contract six months ago (firmware, RDU, OS, etc., versions have changed since, but bug persists):

 

Code:

---

Re:  Mac Retrospect 6.1 Workgroup

Xserve G5 single processor, 2 GB RAM, Mac OS X Server 10.4.8 (pretty vanilla, light load)

Retrospect Workgroup 6.1.126 with RDU 6.1.8.101

Exabyte VXA-2 1x10 1u PacketLoader Plus (SCSI interface)

Exabyte VXA-2 Tape Drive Firmware Rev.: 210E

Exabyte 1x10 1u Autoloader Firmware: A10D

ATTO UL4D DualChannel SCSI, Exabyte is only device on its channel.

ATTO UL4D Firmware Rev.: v1.52

ATTO UL4D Driver Rev.: v4.00

 

This is simply a minor cosmetic bug report for you to pass along to your programmers,

no action is needed on your part. I've noticed it before but haven't reported it before.

Configuration is given above, but I've seen it with Mac OS X Server 10.4.7 as well as

the present 10.4.8, and I saw it with the older ATTO firmware and driver, too (firmware

v1.50, driver v3.60). All firmware, drivers, Retrospect, etc., are up-to-date as of today.

Does not affect operation, cosmetic only.

 

Issue:

If an autoloader slot is empty on an Exabyte 1x10 1u PacketLoader while a drive cleaning

cycle is initiated from the autoloader's front panel while Retrospect is displaying the

Storage Devices window (Configure > Devices), after the cleaning cycle is completed,

Retrospect will show a previously (and still) empty slot as containing a "(cleaning tape)".

 

Repeat by:

Define slot 10 on the autoloader as a cleaning slot. Fill the autoloader with barcoded

tapes, leave at least one slot open. Launch Retrospect, display the Storage Devices

window (Configure > Devices). Observe that the empty slot is shown by Retrospect

as empty. From the front panel of the Exabyte autoloader, initiate a drive cleaning

cycle while Retrospect is displaying the Storage Devices window. Cleaning tape will

move into the drive (and Retrospect will detect it there), and, after completion, will

move back to slot 10. When the cleaning cartridge returns to slot 10, that slot will

show the barcoded cleaning cartridge, but the empty slot will no longer display empty,

but will show "(cleaning tape)". The Exabyte autoloader front panel shows that the

slot is, in fact, empty. Only a cosmetic problem with Retrospect, does not affect

operation. As soon as a barcoded tape is put into the slot, the barcoded tape's

information replaces the erroneous "(cleaning tape)" in Retrospect's display.

Here's the log from Retrospect:

 

Retrospect version 6.1.126

launched at 11/2/2006 3:08 PM

+ Retrospect Driver Update, version 6.1.8.101

3:09:58 PM: Drive "Exabyte VXA DC, SCSI-A:1:0" was cleaned.

 

! 11/2/2006 3:12 PM: Manual erase of tape "2-B 0011"

 

! 11/2/2006 3:12 PM: Manual erase of tape named “2-VXA Set B [011]”

Quit at 11/2/2006 3:18 PM

 

Quit and relaunch of Retrospect does not clear the "(cleaning tape)" in the

Retrospect display for the empty slot.

 

As a comment, prior to the above test, there were two slots empty (slots 2 and

9). When I did a cleaning cycle there, the erroneous "(cleaning tape)" appeared

in slot 9. After putting a barcoded tape into slot 9, the info for the barcoded tape

replaced the erroneous "(cleaning tape)" in slot 9, and slot 2 was still shown as

empty. When I did another cleaning cycle, just to see what happened, this

erroneous "(cleaning tape)" info now appeared in slot 2, which, at the time,

was the only empty slot.

 

Workaround:

Fill up all slots with barcoded tapes. If you then remove a barcoded tape, the

slot returns to the empty information.

 

Cosmetic only, does not affect operation.

---

 

Russ

[jpbuse]

Russ,

 

Wow, a lot of info. Stupid question/answer first. The tape is indeed the proper CLNxxxx barcode label. The A00022 barcoded tape is a good backup tape which ran last night,

 

Retro was happy with it at one time. When I look at the library web gui, it shows the proper tape, in the proper slot, etc. My guess is a cosmetic issue with Retrospect. The library is loaded (24 tapes, including cleaning), and currently there is no tape in the drive. All other backups are/were working just fine.

 

The library is full. Slot 1 is cleaning, slot 2 is the A00022 tape.

 

Retro seems to be reading the element ID and not the slot ID. Retro only shows 23 slots but its a 24 port including the cleaning tape.

 

Jason

[rhwalker]

Then I don't understand. You say:

 

Quote:

---

The tape is indeed the proper CLNxxxx barcode label. The A00022 barcoded tape is a good backup tape which ran last night.

 

Slot 1 is cleaning, slot 2 is the A00022 tape.

 

---

Your second picture is captioned:

 

Quote:

---

Cleaning slot turn off (showing the correct tape barcode):

 

---

Yet there is no CLNxxxxx barcode in any slot, and A00022 is always shown in slot 1.

 

Confused,

 

Russ

[jpbuse]

I guess that is one reason I am really confused. There is a cleaning tape (CLNxxxxx) physically in slot #1 of the tape library. Slot #2 of the tape lib as the A00022 tape in it. So, why does screw it up? It looks like Retro doesn't see the cleaning tape CLNxxxx at all. Maybe I have configured something wrong but I thought in Retro I would simply click on a slot # and designate that as the cleaning slot. Below is a dump from the tape library's inventory system... you can see the cleaning tape is in Slot #1 and the A00022 is in slot #2 (element ID #1). Retrospect is showing 23 of 24 slots. I'm guessing Retro recognizes the CLN tape and doesn't count it as a usable slot. I do have the Exabyte drive setup with a dedicated cleaning slot, meaning I have programmed the tapelib to use a fixed slot #1 and autoclean when needed. Perhaps that would affect how Retro sees it and maybe I don't need to designate a cleaning slot within Retro itself.

 

I ran a test. I disabled the fixed cleaning slot in Retrospect and ran my backup that I was having issues with (to A00022) and sure enough, it worked. It grabbed the proper tape and started to write data to it.

 

Element ID Type: Number Status (POVA) Label valid Label

97 Robot Empty (1010) 1

Clean Occupied (1111) 1 CLNA01

1 Slot: 2 Occupied (1111) 1 A00022

2 Slot: 3 Occupied (1111) 1 A00027

3 Slot: 4 Occupied (1111) 1 A00004

4 Slot: 5 Occupied (1111) 1 A00046

5 Slot: 6 Occupied (1111) 1 A00047

6 Slot: 7 Occupied (1111) 1 A00003

7 Slot: 8 Occupied (1111) 1 A00020

8 Slot: 9 Occupied (1111) 1 A00039

9 Slot: 10 Occupied (1111) 1 A00016

10 Slot: 11 Occupied (1111) 1 A00001

11 Slot: 12 Occupied (1111) 1 A00023

12 Slot: 13 Occupied (1111) 1 A00030

13 Slot: 14 Occupied (1111) 1 A00017

14 Slot: 15 Occupied (1111) 1 A00018

15 Slot: 16 Occupied (1111) 1 A00031

16 Slot: 17 Occupied (1111) 1 A00033

17 Slot: 18 Occupied (1111) 1 A00034

18 Slot: 19 Occupied (1111) 1 A00036

19 Slot: 20 Occupied (1111) 1 A00035

20 Slot: 21 Occupied (1111) 1 A00042

21 Slot: 22 Occupied (1111) 1 A00043

22 Slot: 23 Occupied (1111) 1 A00045

23 Slot: 24 Occupied (1111) 1 A00037

E/E Port In Process (1000) 1

81 Drive: 1 Empty (1011) 1

 

Quote:

---

Then I don't understand. You say:

 

Quote:

---

The tape is indeed the proper CLNxxxx barcode label. The A00022 barcoded tape is a good backup tape which ran last night.

 

Slot 1 is cleaning, slot 2 is the A00022 tape.

 

---

Your second picture is captioned:

 

Quote:

---

Cleaning slot turn off (showing the correct tape barcode):

 

---

Yet there is no CLNxxxxx barcode in any slot, and A00022 is always shown in slot 1.

 

Confused,

 

Russ

 

---

[rhwalker]

Ah. I see. Appears to be confusion between Retrospect's designation of a cleaning slot and the autoloader's designation of a cleaning slot. Perhaps I am seeing something different because my cleaning slot is designated as the highest slot. Here is what Retrospect shows for devices:

Code:

---

Exabyte Library, SCSI-A:0:0

Exabyte VXA DC, SCSI-A:1:0

(Empty)

Library slots 1 to 10

1: [1-A 0011] 1-VXA Set A [011]

2: [2-A 0011] 2-VXA Set A [011]

3: [1-A 0012] 1-VXA Set A [012]

4: [2-A 0012] 2-VXA Set A [012]

5: [1-A 0009] 1-VXA Set A [009]

6: [1-A 0010] 1-VXA Set A [010]

7: [2-A 0010] 2-VXA Set A [010]

8: [1-B 0012] 1-VXA Set A [011]

9: [2-B 0012] 2-VXA Set B [012]

10: [CLNA0002] (Cleaning tape)

---

Here's the output from libtool:

Code:

---

mail:~/Documents/Exabyte/packetloader_firmware admin$ ./libtool Changer1 -l

libTool V1.79 - © 2002-2006 Exabyte Corp.

 

Changer1 - Device recognised as Exabyte PacketLoader 1U

Changer1 - Retrieving library content...

Slot 1 [Occupied, Barcode = 1-A 0011 ]

Slot 2 [Occupied, Barcode = 2-A 0011 ]

Slot 3 [Occupied, Barcode = 1-A 0012 ]

Slot 4 [Occupied, Barcode = 2-A 0012 ]

Slot 5 [Occupied, Barcode = 1-A 0009 ]

Slot 6 [Occupied, Barcode = 1-A 0010 ]

Slot 7 [Occupied, Barcode = 2-A 0010 ]

Slot 8 [Occupied, Barcode = 1-B 0012 ]

Slot 9 [Occupied, Barcode = 2-B 0012 ]

Slot 10 [Occupied, Barcode = CLNA0002 ], [Cleaning Slot]

Drive 1 [Empty]

All drives are reported empty

 

Done

mail:~/Documents/Exabyte/packetloader_firmware admin$

---

Note that I've also told the autoloader that slot 10 is the cleaning slot.

 

Here is the inventory output from the web admin page for the autoloader:

Code:

---

Inventory

Index What POVAR Label valid Label

1 Slot 11110 1 1-A 0011

2 Slot 11110 1 2-A 0011

3 Slot 11110 1 1-A 0012

4 Slot 11110 1 2-A 0012

5 Slot 11110 1 1-A 0009

6 Slot 11110 1 1-A 0010

7 Slot 11110 1 2-A 0010

8 Slot 11110 1 1-B 0012

9 Slot 11110 1 2-B 0012

10 Clean 11110 1 CLNA0002

82 Drive 10110 1

86 Robot 10100 1

---

Russ