What is new in SPD 3.4?

 Hello,

 are there any new features in Version 3.4? Or is it mainly a bugfix release?

Comments

STC-Nate

Re: What is new in SPD 3.4?

3.4 is a maintenance release.  Bug fixes.  There are a couple of odd new features, but they're very minor.  One new feature is that you'll find that whenever a new image file is created, a text file containing the MD5 sum/hash for that file is created next to it.  This MD5 is calculated using the original data before it is even written to the image file, so these MD5 values are of maximum quality (they will even reveal if any of the writes to create the image file itself failed to do their job).  Also, of course, this MD5 allows you to verify images using third-party utilities, which can be helpful.

Another feature which is present in 3.4, although not exposed in the GUI, is the ability to ignore read failures (caused when the hard disk itself is failing) and continue a backup.  When this feature is used, the backup will backup zeroes for those sectors which could not be successfully read.  To enable this option, you create a system environment variable named SP_SBVOL_IGNORE_READ_ERRORS with the value 1 and then restart (stop and start) the "ShadowProtect Service" (ShadowProtectSvc.exe) for the change to take effect.  Oh, in 3.4 we also added support for Portuguese (Portugal variation).

If you aren't experiencing any issues with 3.3 then there's no hugely compelling reason for you to upgrade to 3.4.  It's mostly for those folks who are experiencing issues.

Mr. Nice Guy

Re: What is new in SPD 3.4?

 Thanks, Nate :)

Pleonasm

Re: What is new in SPD 3.4?

Nate, if a user is encrypting the backup image, is the following statement still accurate?

 

“This MD5 is calculated using the original data before it is even written to the image file, so these MD5 values are of maximum quality (they will even reveal if any of the writes to create the image file itself failed to do their job).  Also, of course, this MD5 allows you to verify images using third-party utilities, which can be helpful.”

STC-Nate

Re: What is new in SPD 3.4?

Yes, it is still correct.  The MD5 sum is computed using the data which is about to be written to the new image file, before it is actually passed to the OS for the file write operation.  This means that the MD5 sum is the most accurate hash for this type of scenario.  If the data write operation (write to the new image file) fails to actually write the data and doesn't return a failure then this MD5 sum will even catch such an error.  And of course, the data is already encrypted and compressed before it is used in the MD5 sum calculation.

Bulldog

Re: What is new in SPD 3.4?

 1. Is the MD5 calculated from about-to-be-written data verified after the image is actually written? Or is it just stored?

 2. Without the SP_SBVOL_IGNORE_READ_ERRORS environment variable, what does SP Desktop normally do after a read failure?

STC-Nate

Re: What is new in SPD 3.4?

I can't figure out your question 1 (forgive me) so instead I'll just try to reword my previous answer and cross my fingers in the hope that it'll somehow answer whatever question it is that you have.

Here's the basic sequence of steps as the data is read from the source and then written to the destination, in relation to the MD5 hash calculation operation.

  1. A block of data, to be backed up, is read from the source volume
  2. The block of data is optionally compressed
  3. The block of data is optionally encrypted
  4. The MD4SUM is updated based on the content of the block of data
  5. The block of data is written to the new image file

Steps 1-5 are repeated until all blocks from source are written to target. The target file is then closed (as it is complete) and then another new file (a text file) is created with the same name as the new image file but with a .MD5 extension on the filename rather than .spf or .spi.  This text file contains the MD5 sum represented as a string of text.

Your second question is unrelated to the MD5 stuff.  I just want to clarify this so that readers aren't confused by the two.  On this unrelated topic, if SP_SBVOL_IGNORE_READ_ERRORS is not defined, then during a backup if the source disk fails a read request then the failure is logged in the job's detailed log and the backup job fails.  It will reattempt another backup the next time its schedule dictates (and that backup may very well fail too).  If you have scheduled email notification on failure, then you will be notified by email on any such failures.

 

Bulldog

Re: What is new in SPD 3.4?

[quote user="Nate"]

I can't figure out your question 1 (forgive me) so instead I'll just try to reword my previous answer and cross my fingers in the hope that it'll somehow answer whatever question it is that you have...

[/quote]

Thank you. That answers the question I had in my head. :-)

[quote user="Nate"]

....if SP_SBVOL_IGNORE_READ_ERRORS is not defined, then during a backup if the source disk fails a read request then the failure is logged in the job's detailed log and the backup job fails...

 [/quote] 

The new environment variable seems like such a great feature I wonder why it's not exposed in the GUI, perhaps with a toggle. If the read fails in just one or a handful of sectors, and there's a chance of recovering data from the remaining sectors, maybe I'd like to give it a shot.

 Thanks for your help.

STC-Nate

Re: What is new in SPD 3.4?

It's not in the GUI mainly because it was not intended as an exposed feature in this release (even though it was implemented in core components).  However, enough people have needed it that we decided to simply reveal the workaround.  Mostly this is a matter of timing.  Every time you change the GUI you have to get a lot of people involved to translate your changes.  If you're close to release, you really don't want to mess with you GUI because orchestrating the translation will delay the release.

Terms and Conditions of Use - Privacy Policy - Cookies