Discussions specific to the Microsoft Hyper-V hypervisor
Post Reply
StephanF
Enthusiast
Posts: 54
Liked: 16 times
Joined: Mar 26, 2015 1:15 pm
Contact:

Understanding guest file indexing (possible feature request)

Post by StephanF » Sep 11, 2017 2:38 pm

Hi,

I have a question regarding the design of guest file indexing. I know that this is needed to be able to search for files in Enterprise Manager.

To my observation it is working the way that during backup a service is spawned inside the VM that indexes all files. To me this has two drawbacks:
1. This generated load (CPU and storage) on the production VM.
2. file system indexing can take much longer than backup itself. This increases job runtime.
Additionally I don't know if there is a mechanic that keeps backup and index in sync. Backup is taken from a snapshot while indexing service is running in the production VM itself. What happens if files are deleted in the VM after the backup snapshot was taken but before the indexing service saw them? Are they missing from the index?

In my oppinion a post backup index approach would be much more suitable. Veeam is able to mount volumes from the backup. Why not doing the backup from the production environment as fast as possible and after that mount the volumes on the backup server from the backup repository and create the file system indexes? This could work just like SureBackup does. Maybe as an own scheduled job in times with low load.
Wouldn't this reduce the load in production and have some advantages?

You can take this as a feature request or convince me that I am wrong. I just want to understand it.

Thanks
Stephan

foggy
Veeam Software
Posts: 18291
Liked: 1568 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by foggy » Sep 11, 2017 4:43 pm

Hi Stephan, you're right that guest file system indexing is performed by a runtime process deployed inside the VM. However, the indexing procedure uses the VSS snapshot, not the VM guest OS itself, and is carried out in parallel with the backup procedure itself, so as a result, the created file index exactly reflects the state of the backed up VM. Indexing speed depends on the number of files, however, typically is pretty fast due to getting data directly from MFT rather than by scanning the entire file system.

StephanF
Enthusiast
Posts: 54
Liked: 16 times
Joined: Mar 26, 2015 1:15 pm
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by StephanF » Sep 12, 2017 11:54 am

Thank you for the explanation. It is really good to know that guest file system indexing is using the VSS snapshot. This eliminates one of my fears that there could be inconsistent data sets.
Speed however depends on your point of view. We have a file server with several million of small files. Daily incremental backup takes about 30 minutes while guest file indexing takes 3 hours. I don't is this is fast ow slow. At least it is not an issue for us. So everything is ok on my side.

In my opinion the possibility to do a post-process file indexing from existing backups would have some benefits. But I admit there is no common issue for making this a high priority feature request.
Maybe some other people will notice this threat and have an opinion. Maybe they don't.

For me my questions have been answered. Thank you.

foggy
Veeam Software
Posts: 18291
Liked: 1568 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by foggy » Sep 12, 2017 11:59 am

Indeed, in case of millions of files it could take a while to get, parse and write their descriptions. Anyway, your feedback is appreciated and request is accepted, thanks!

Nick-SAC
Enthusiast
Posts: 40
Liked: 5 times
Joined: Oct 27, 2017 5:42 pm
Full Name: Nick
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by Nick-SAC » Apr 09, 2018 3:42 pm

Hello,

In: Backup Job > Guest Processing > Guest File System Indexing > Indexing (Selecting 1 of 3 VMs) > Edit > Windows Indexing :
The “Index only the following folders > Add...” dialog, only lists the Host (not VM Guest) Folders.

My desire is to index only specific Volumes and/or Folders on the Guest VM.

What am I missing or just not understanding?

Thanks,
Nick

csydas
Expert
Posts: 193
Liked: 46 times
Joined: Jan 16, 2018 5:14 pm
Full Name: Harvey Carel
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by csydas » Apr 14, 2018 8:52 am

Hi Nick,

We hit this before, and it's just a matter of a not clearly explained thing.

Type in the folder paths you want to include - The Veeam Console isn't aware of what's inside your production VMs, much less the at that stage. The assumption seems to be that most of what people want to exclude at the most common directories that exist in Windows, so there's a browser based on the Veeam's host system layout.

So if you want D:\ included or excluded, just type it in and press Add. or just D:\vacation_pics, likewise. It's funny, but when you remember there isn't a persistent agent from Veeam, it starts to make more sense.

Nick-SAC
Enthusiast
Posts: 40
Liked: 5 times
Joined: Oct 27, 2017 5:42 pm
Full Name: Nick
Contact:

Re: Understanding guest file indexing (possible feature requ

Post by Nick-SAC » Apr 14, 2018 2:46 pm

Thank you Harvey,

I just plugged in what I want and I’ll watch the next backup... (I’m still in testing & eval mode).

I’d suggest that in lieu of being able to Browse the Guest FS; just a little [Manually Enter What You Want] notation on that dialog would help avoid the confusion...

Thanks again,
Nick

Post Reply

Who is online

Users browsing this forum: No registered users and 8 guests