[Tim Jones]

Yes – what Bret said :). I guess we were posting at the same time.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

One method is to follow these steps:

Create a New Library on your backup drive
Name the Event properly for your backup copy
Select the project that you want to offline and backup
Drag it into the new Event in the new library

I like to include the optimized media, but I leave the proxy media out of the new copy. Note that this doesn’t copy linked files, just the alias entries (references).

Select the Library and select “Reveal in Finder”.
Select the Library there and Right click and select “Show Bundle Contents”

If all looks correct, you can now delete the original project.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

One more thought about the formatting step –

Try formatting with this command sequence:
mkltfs -f -d 0 -s RR0001 -r "size=2M/name=Manifest.xml" -n "Red Rock Deliverables"
(of course, change the serial number and volume name to suit your needs..).

Now mount that new volume and copy the files that are currently erroring.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

One more thing for details –

When you execute that ltfs command to mount the volume, there are a lot of lines spit out in the terminal window that look like this:
LTFS14000I LTFS starting, LTFS version 2.0.0, log level 2
LTFS14058I LTFS Format Specification version 2.0.0
LTFS14063I Sync type is "time", Sync time is 300 sec
LTFS20013I Drive type is HP LTO6, serial number is XXXXXXXXXX
LTFS17160I Maximum device block size is 524288
LTFS11005I Mounting the volume
LTFS12015I Attempting to load the medium
LTFS11031I Volume mounted successfully
LTFS14029I Ready to receive filesystem requests

Change your mount command to spit out a bit more detail and then try to copy the failing files over to the tape:
ltfs /Volumes/SampleLTFSMount/ -o trace,eject,volname="Sample LTFS Mount",devname=0
LTFS20028D Found device class 0x5AC92550 at index 0
LTFS20010D SCSI request: [ 12 00 00 00 F0 00 ] Requested length=240
LTFS20011D SCSI outcome: Driver status=0x00 SCSI status=0x00 Actual length=96
LTFS20084D Drive identification is 'Ultrium 6-SCSI '
LTFS20010D SCSI request: [ 12 01 80 00 20 00 ] Requested length=32
LTFS20011D SCSI outcome: Driver status=0x00 SCSI status=0x00 Actual length=14
...

After you copy a file that normally fails, scroll through the output until you find the filename that you tried to copy and you’ll see output that looks like this:
LTFS20057D Backend capacity part0: (36750, 36750)
LTFS20057D Backend capacity part1: (1384552, 1396529)
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14051D FUSE getxattr '/' (name='com.apple.FinderInfo')
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14040D FUSE create file '/Minimus the Poet Final.fcpxml'
LTFS20010D SCSI request: [ 00 00 00 00 00 00 ] Requested length=0
LTFS20011D SCSI outcome: Driver status=0x00 SCSI status=0x00 Actual length=0
LTFS14030D FUSE fgetattr for '/Minimus the Poet Final.fcpxml'
LTFS14031D FUSE getattr for '/._Minimus the Poet Final.fcpxml'
LTFS14031D FUSE getattr for '/Minimus the Poet Final.fcpxml'
LTFS14037D FUSE flush '/Minimus the Poet Final.fcpxml'
LTFS14035D FUSE release file '/Minimus the Poet Final.fcpxml'

Walk through that text to find the actual error condition that is being reported. Maybe that will give us more detail on why it’s failing.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

First, the 2.3.8 update was created for a no longer supported version of BRU PE so that it would run on Mountain Lion properly. That is truly the big difference between 2.3.6 and 2.3.8. Even though the 2.3 family tree is end of life’d and no longer supported directly, we at least tried to give existing 2.3 users a version that still worked on 10.8. Try and get that from Avid 🙂 .

As to the corruption, support tells me that they have no support ticket about this from you (support.bru.com). If you had opened one when this happened, they would have provided the following info:

When this happens (a binary application corruption), it usually means that there is something wrong in the application binary bundle (OS X apps are far more than a simple single file) . The solution is to delete the application package and reinstall. This is true for ANY application that you install on OS X, not just BRU PE. In the case or BRU PE, simply delete the “BRU PE.app” bundle in your Applications folder, empty the trash, and then rerun those command line instructions by using copy and paste.
sudo /usr/bin/true
gunzip -c ~/Downloads/BRU_PE_2.3.8.737_GUI-Only.bgz | sudo bru -xvf - -ua
The first line just sets up administrative privileges for the second command. The second command simply extracts the compressed bruball and restores the files to where they belong.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

With RAID-6, you can lose 2 drives within the array before your data is gone. On the other hand, with 2 RAID-5’s mirrored, you could lose an entire array and one drive in the second array. Given an unlimited budget, I’d run a stack of four 4TB-based 8 drive arrays – each as RAID-6. I would then stripe two arrays into a pair and then mirror each striped pair.

More redundancy is ALWAYS better. Your real limitation in this realm is budget.

However, this would fall into the realm of serious overkill unless you were forced to use questionable drives in the arrays.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

In the days of LTO-4 and earlier, the Full height drives provided an average of 50% better throughput than the half height versions of the same drive. This was because the mechanicals were larger and more robust and there could be more sensors installed in the tape path to monitor and control tape speed and positioning. Since LTO-5’s second update, the mechanical differentiation has become NIL and the actual tape path components are far more “exact” than in prior versions.

You are correct about the ratio of full height to half height. Even HPs’ internal numbers (were they public) would probably match ours of over 22:1.

In the current (2014) LTO-5 and LTO-6 world, we (TOLIS Group) don’t even sell the full height devices except for use in 3rd party libraries that were only designed for full height drives. They bring absolutely nothing to the table. In fact, we even designed and had built our own enclosure that uses the HH device and provides space for either 2 tape drives or a drive and a 2.5″x4 drive disk-cage for more reliable cooling and MUCH quieter desktop operations.

Save your money and buy half height. And, remember that an LTO-5 drive will read / restore from LTO-3 tapes. so let that LTO-3 RIP.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

It is VERY important to note that BackBlaze admits to using only the CHEAPEST drives possible. While there is a market for these drives, it’s also important to understand that there is a reason for the pricing delta between a consumer-grade drive (all of the units mentioned by BackBlaze are consumer / desktop drives) and Enterprise-grade drives. Our experience with the Seagate and WD desktop drives is very similar to the numbers presented by BB. However, our experience with Seagate’s ES class 3.5″ and 2.5″ drives is very different – in the sub-1% range over thousands of customer-placed drives.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!

[Tim Jones]

While it would be great if there could be a cut and dried answer, there are a few questions here that you should consider –

[Tim Jones]

In that case, my aside answered your query – HP and Fujifilm have proven to be the most reliable and least abrasive of the brands and are all that we recommend to our customers.

Tim
—
Tim Jones
CTO – TOLIS Group, Inc.
https://www.productionbackup.com
BRU … because it’s the RESTORE that matters!