Backup image being split into multiple files?
Hi, since October, TIH has frequently been splitting my C and D drive backup images into 2 - 7 random sized files instead of a single file. When it does this, it adds "_v1, _v2, _v3" etc. to the file name for each segment. The segments vary from 90MB to 50+GB in size.
This is a Windows 10 desktop and the problem never occured prior to October. Originally running TIH 2015 then 2016 since August. Updated to 2017 in December with same issue. Total image size is @50GB for C and @230GB for D. Backup is to a 3TB Netgear ReadyNAS RAID 5 over a certified performance hardwired gigabet home network. No other network traffic or NAS use occurs when the backups run. Backup routine is always Full for C and custom Full/6 Incrementals for D. This problem occurs most of the time but not always.
(As a side question: how do you manage backup images in TIH?)
Thanks!
Hi, Thanks for your reply
I have 1.8TB dedicated on the NAS for these two routines. The C backup keeps 6 Full images (@45GB each). The D keeps 1 full set (1 Full + 6 Incrementals). The NAS alerts me when it's down to 300GB available and I regularly monitor it. When TIH fails to automatically delete a prior backup set, I do it manually. I've run these same two routines on the same equipment (desktop, NAS, switches, cables) for 3 years and never had this issue until October. Over time, the C file size has increased by @5GB and the Data drive by about 30GB. I don't know how raided drives deal with fragmentation, but the destination drive capacity hasn't been reduced, so it shouldn't be the issue.
Any other thoughts, like validation or "automatic" file size setting?
(Tonight I'm going to run the D drive routine on a new, empty direct-connected 2TB USB external drive & see if it does the same issue)
Does your NAS report SMART errors or information to you? If yes have you viewed them? Do you see any errors like CRC errors?
If you do I suggest that you shutdown the NAS, reseat all drives and drive data cables everywhere possible.
Your problem might be due to interuption in the data stream from your PC to the NAS.
Bobbo's points are good as well. You might have a disk in the array going bad and that could be causing the problem as well. Best to check your hardware.
Thanks! Towards the end of December, one of the drives started rapidly increasing reallocated sectors, so I replaced it. SMART on all drives has been normal since, but still have the issue.
When yoiu replaced the failing drive, I'm sure the RAID automatically rebult, mirroring/stiping the data as it was all written to the other drives. That's fine - that' s how it works.
However, even if you have free space on the drive, the available block sizes will get smaller and smaller as the disk fragments with different read/writes. We don't normally see this as an issue when drives are not fragmneted or when they're new, but we also mostly never see this because we typically don't deal with very large single files. Backup programs are more likely to show you this because if you take the default sizes, they usually try and make 1 big file or as many files as it takes, using the largest amount of free contiguous space. Consider the next single largest file on your computer - if you have blu-ray images or .iso's, that's probably the largest at about 24GB each and most people don't have a single file that's larger than 4.7GB (DVD .iso - although I've seen people let their Outlook .PST's get wild, but generally not more than 10GB).
wincontig is a program to specifically move all files into contiguous space - it's like defrag for files. If you fun it on an entire drive (or parition), then techncially, you would have a big block to use again. I wouldn't run it against a NAS though - I only use it for smaller flash drives where it's necessary to put make .iso's contiguous so they load into memory correctly.
If the problem is capacity, block sizes, etc - I would expect the problem to become progressively worse, but it is random. (Keep in mind no other variables (data changes) exist on the destination drive - it's exclusive for these backups)
Example: 12/27 - C Full - 1 file, 32GB
1/1 - System Full - 5 files, @34.5GB
1/3 - C Full - 3 files, @32GB
1/8 - System Full - 1 file, 34.6GB
Prior: C Full made single file images on 11/15 & 11/22 - multi file on 11/29, 12/6 & 12/20 // System Full single file images on 11/6, 11/13, 12/18 & 12/25 - no other multi-file
Note: This pattern began mid-October. Almost all D drive backups since then, both full & incremental were multi-file images. This problem never occurred during the two+ years prior (I recently stopped running D drive backups due to a Credentials issue with build 5554)
I would expect that your backup to the local 2TB drive will show a single full backup image file, and that the issues reported for the NAS are as suggested, a matter of fragmentation.
See web document: ReadyNAS OS 6: How to Enable and Configure Volume Maintenance from the Netgear KB site which deals with such fragmentation and how to deal with this on your NAS.
See also: ReadyNAS OS 6: An overview of relevant scheduled maintenance tasks available where it has the following paragraphs:
How is fragmentation different on ReadyNAS devices compared to traditional storage devices?
Basically, fragmentation occurs on ReadyNAS devices as it would on most mass storage devices. However, there is one difference with ReadyNAS devices. When the bitrot protection functionality is enabled the ReadyNAS device moves even further away from traditional storage devices in the way it handles files. One of the effects is that new files are stored only on genuinely free blocks in contrast to the options mentioned above. This leads to a slightly higher rate of fragmentation
Should I use defragmentation regularly?
Fragmentation produces a considerable amount of data overhead, so defragmenting a volume can lead to an increase in free space available. Read performance will most likely also be increased when using magnetic hard disc drives.
However, in some cases defragmentation can also lead to reduced free space available. This again is connected to the new methods implemented in the ReadyNAS file system. More precisely, defragmentation can lead to losing the advantage of these methods as you might be losing the advantage of zero size snapshots for existing snapshots. Basically, snapshots are disconnected from the data they are pointing to and are therefore manifested on the storage device taking up additional disc space. Finally note that defragmentation puts stress on the hard discs.
Taking all into account, defragmentation is advisable at sparse regular intervals.
As a test, set the default TIB size for the backup to, say, 10GB and see if it continues to break them up even smaller or if it likes that size.
Another thought... VSS. VSS has to have room on the original drive (C: drive) to cache all of the data before writing it. This may be where the problem is actually occuring - the local OS drive and not the NAS. How much free space do you have on your C: drive parition? What is the system disk space usage for "system protection" set to on the C: parition as well?
HardToThrill, if you have updated to the build 6616 ("Acronis True Image 2017 New Generation"), we'd be interested to know if the issue with split backups persisted.
To see if the issue is indeed in VSS, as Bobbo suggested, you can change the snapshotting method from VSS to Acronis snapshot under Options - Advanced - Performance, save changes, run the backup and see if it is created as a single file or not.
It would also make sense to try disabling any kind of offloading in the network adapter's advanced properties. Also check if any updates were installed for the network card driver, maybe try to revert to a previos version driver. Often backups are split when a network card tries to remove the load from CPU (offload setting is ON), but sometimes (not always) is unable to do that properly for any reason (issues with new drivers, hardware failures). Then backups become split randomly when saved over network.
Regards,
Slava