The support forum

Poll on VSS Issues...

Deipotent :

Jun 27, 2013

I have already been corresponding with Alex about regular VSS failures while backing up files in use, and wanted to see if anyone else was getting VSS errors when files are in use ?

Trying to see if my regular failures are the norm, or unexpected.

Can you post your OS, total size of source partition, and free space on source partition.

pjb :

Jun 27, 2013

With another backup program I used in the past, they discussed possible causes of VSS failures:
http://www.2brightsparks.com/help/index.php?action=kb&article=228

I live image a server I am supporting and have not had issues related to VSS since I shut down the mysql database while imaging with that program.  

Although as I revisit the link, I was having unexplained frequent timing-out issues with a software vendor's built-in database backup program (which does an sql dump of an active mysql database) from the EMR vendor package the office uses.  Maybe the timed-out issue is due to VSS in that case so I need to investigate further.

JacobMoreagles :

Jun 28, 2013

I'm getting VSS errors as well. In fact, I'm not certain it's ever worked for me. The files I have in use are sql databases, and they get backed up by other mechanisms, so it's only at the off chance I'd be using some other file that i'd need it.
It should be said however, VSS not working is far from a dealbreaker for me, the rest of the program works awesome.

This is backing up a 3tb array on windows server 2008r2. The array has ~700gb free, and the total backup size is ~300,000 files and it's backing up onto 2 3tb usb3 drives nightly.

Deipotent :

Sep 04, 2013

I came across a couple of people who mentioned that RAM Disk's can causes problems for VSS. I have since updated my RAM Disk software, which supports VSS and has an option for hard disk emulation, and in my limited initial testing VSS snapshot creation is not failing any more.

NOTE: neither the job source or destination were a ram disk, so it looks like other device drivers (un-related to the actual src/dest) in the storage stack can cause problems for VSS.

Alex Pankratov :

Sep 04, 2013

Good to know, thanks for update.

Deipotent :

Sep 04, 2013

I have since updated my RAM Disk software, which supports VSS and has an option for hard disk emulation


"... which <b>*now*</b> supports VSS and..."

Ramley :

Jan 28, 2014

It's my first time writing to this forum. Then first of all congratulations Alex, for producing so powerful and well-designed SW while keeping it that simple, unobtrusive and intuitive. I have used many backup utilities in the last 15 years and this is something that really states apart! I'm using it to keep a real-time mirror of my daily work into an encrypted SD card, getting ready for frequent hard disk failures (twice last year with the shitty SSD disks that my company forces me to use).

I saw several big improvements from previous Bvckup 1 public beat, including VSS for Windows 7 on my X64 architecture. But I see now a -minor- issue when making backup of an external, 16 GB FAT32 formatted USB drive. VSS fails with this error in Bvckup error log "IVssBackupComponents::AddToSnapshotSet() failed with 8004230c". Looking into Microsoft pages this means that "The value of the ProviderId parameter is GUID_NULL and no VSS provider indicates that it supports the specified volume or remote file share". Excuse my ignorance, does it mean that it is not possible to take a VSS snapshot from an externally attached USB drive? After that error message, Bvckup2 says "Snapshot was NOT created" and it continues without VSS, which in fact is not a big deal for me for backing up external devices.

By the way, in this link they mention VSS failures due to attached USB drives that use 4k physical sectors, mostly the norm for today's drives higher that 1 TB. That's not my case but it could be the reason for other people's issues.

http://support.unitrends.com/ikm/questions.php?questionid=1198

Deipotent :

Jan 28, 2014

I'm sure Alex will add to this but I believe VSS requires the source to be formatted as NTFS, so it seems like you've found a bug where Bvckup is attempting to create a VSS snapshot on a FAT32 partition.

Alex Pankratov :

Jan 28, 2014

@Ramley - Thanks for the kind words, really happy to hear that!

I'm using it to keep a real-time mirror of my daily work into an encrypted SD card, getting ready for frequent hard disk failures.


I would consider also adding a daily and, perhaps, weekly backup. Just in case the primary drive goes down gradually, first corrupting data here and there and only then dying.

8004230c


0x8004230c means "Shadow copying is not supported on this volume" and as @Deipotent said it's due to the drive being FAT32 rather than due to it being external. I'll add a better message around this error, but that's all I can do here :)

Ramley :

Jan 28, 2014

Great Alex & Deipotent, thanks for the fast response!

Deipotent :

Jan 28, 2014

Surely this is a bug though @Alex, as Bvckup knows the type of the file system is FAT32, so why is it even trying to create a VSS snapshot ?

Alex Pankratov :

Jan 29, 2014

Not really a bug, no. In theory it's possible to have a FAT32 shadow copying provider, so deciding on whether it is supported based on the file system type is not fool-proof. It should instead be using [1] and I have made this change.

[1] http://msdn.microsoft.com/en-us/library/aa382692%28v=vs.85%29.aspx

adlerweb :

Mar 07, 2014

Just to document my "usual" VSS issue:

OS is a Win2k8R2 Cluster, Source partition does not really matter here. Bvckup triggers all VSS writers including the cluster database which will not finish in time and generate warning 1544 inside the cluster event log ("The backup operation for the cluster configuration data has been canceled. The cluster Volume Shadow Copy Service (VSS) writer received an abort request.").

vssadmin shows

Verfassername: "Cluster Database"
   Verfasserkennung: {41e12264-35d8-479b-8e5c-9b23d1dad37e}
   Verfasserinstanzkennung: {917178cf-ad9b-4d2f-bc41-01a5121c9dfc
   Status: [9] Fehler
   Letzter Fehler: Zeit├╝berschreitung
(last line translates to "Last error: Timeout")

Since i don't really care about the cluster database during bvckup snapshots and the events are easy to filter its not really a problem for me.

Alex Pankratov :

Mar 07, 2014

Thanks. I need to make another pass over VSS part, it does indeed uses blanket approach right now and it could be a bit more selective.

Alex Pankratov :

Jun 26, 2014

To circle back to this issue. I just made a couple of changes to the VSS handling.

1. It now logs all writers and components involved in making a snapshot. Like so - https://bvckup2.com/wip/vss-writers.png

2. It now explicitly includes all optional components.

Previously the code was absolutely oblivious to the existence of writers and (selectable/optional) writer components, so it effectively made snapshots with a default set of writers.

Before the next release (R69) is out, I will also add support for suppressing specific writers on per-backup basis, just in case if someone needs a more granular control over the process.

If anyone is willing to test an interim release, let me know. I have it all packaged up and ready to go.

highend :

Aug 16, 2014

I have a major issue with VSS as well.

- Windows Server 2012 R2 Datacenter
- Bvckup2 v70 running as Administrator (as the window title says)

Before any backup:
vssadmin list writers
All existing writers:
Status: [1] Stabil
Letzter Fehler: Kein Fehler

After a VSS enabled backup to a archlinux driven Pogoplug device
Differences:
Verfassername: "System Writer"
Verfassername: "Microsoft Hyper-V VSS Writer"
Verfassername: "ASR Writer"
Verfassername: "COM+ REGDB Writer"
Verfassername: "Registry Writer"
All show:
Status: [8] Fehler
Letzter Fehler: Inkonsistente Schattenkopie

Great... :(

Now I try to backup to the drive U:\ which is a mounted truecrypt partition from an external USB3 device:

Verfassername: "System Writer"
Verfassername: "Microsoft Hyper-V VSS Writer"
Verfassername: "ASR Writer"
Verfassername: "Registry Writer"
Verfassername: "Shadow Copy Optimization Writer"
Verfassername: "COM+ REGDB Writer"
Verfassername: "WMI Writer"

For all these writers:
Status: [7] Fehler
Letzter Fehler: Zeitueberschreitung

...

Alex Pankratov :

Aug 16, 2014

Do you get the same results if you try and use any other VSS-enabled utility, e.g. hobocopy?

The only variable that is available for tweaking in Bvckup2 is the list of optional writers - you can suppress specific ones from being included into the snapshot. But then if excluding an optional writer would help resolving this issue, then it would imply that the writer itself is buggy.

I mean... from app's' standpoint there's a Microsoft-prescribed way of interacting with VSS and there are no if's except for optional writers. If this doesn't work or if it messes up the VSS state, then there's not much I can do on my end. The problem would seem to lie with the VSS itself.

highend :

Aug 16, 2014

No, I've just tried the same with SyncBackSE

It copied the parent.lock file from the running Firefox instance without problems (via VSS) and the status for all writers after the backup is still:

Status: [1] Stabil
Letzter Fehler: Kein Fehler

Alex Pankratov :

Aug 16, 2014

OK, I just sent you an email with some questions.

New topic

Create
Made by Pipemetrics in Switzerland
Support


Follow
Twitter
Dev blog
Miscellanea Press resources
Testimonials
On robocopy
Company
Imprint

Legal Terms
Privacy