Page 1 of 1

Data corruption error?

Posted: Thu Aug 01, 2019 11:45 pm
by zBeeble
I'm a software engineer. I'm also a systems engineer. I have some idea of what it will take to find and nail this. I do file this bug lightly.

Here's the situation. My windows 10 "workstation" (a higher than "pro" version of windows 10) has the following configuration:

System Summary:
CPU: Threadripper 1900
RAM: 64G (not ECC, so blame that if you like, but...)
Drive C: 2T Samsung 970 nvme SSD using NTFS
Drive D: 4x 4T WD Red using ReFS
Drive F: another ntfs SSD (doesn't matter, not used, only listed for completeness).
Drive not mounted: WD Blue nvme 1T

PrimoCache configuration:
Cached Partition: D
L1 cache: 8G
L2 cache: the unmounted WD Blue SSD.
Defer write enabled (10ms)

Software: Acronis True Image 2019
Acronis is configured to incrementally backup C (the 2T NVME SSD) to a directory on D (the 4T x 4 ReFS). Nightly. This generates a single very large file per backup. Acronis also verifies the written backup each night.

On two occasions now, Acronis has reported errors in the verification. On the first occasion, it seemed the error was in a previous backup (so a backup had passed the nightly check, but checks on subsequent nights failed). On the second occasion, which is ongoing, it has found an error in a nightly backup, redone the backup and the error has gone away. The error looks like this in the log:

2019-08-01T04:08:45:477-04:00 1300 I000B0402: Validate Backup Archive Location: "D:\Acronis\Samsung SSD 970 EVO 2TB 2B2QEXE7_inc_b13_s22_v1.tib"
2019-08-01T04:08:45:492-04:00 1300 I000101F8: Pending operation 4 started: 'Validate Backup Archive'.
2019-08-01T04:15:33:344-04:00 1300 E000101F6: Error 0x101f6: Error while checking the metadata integrity of points in time (PITs) in the archive.
| trace level: error
| line: 0x65b5eb7011094703
| file: c:\bs_hudson\workspace\790\processor\diskadm\da_commit.cpp:348
| function: DaProcessor::CommitImpl::OnDaError
| line: 0x65b5eb7011094703, c:\bs_hudson\workspace\790\processor\diskadm\da_commit.cpp:348, DaProcessor::CommitImpl::OnDaError
| $module: ti_demon_vs_17750

... interestingly... it has always been the "metadata" that is in error.

Let me know how you'd like to proceed.

Re: Data corruption error?

Posted: Fri Aug 02, 2019 3:03 am
by support
Thank you very much for the bug report!
Could you run the command "chkdsk d:" to see if there are errors in the volume D file system?

Re: Data corruption error?

Posted: Fri Aug 02, 2019 3:06 pm
by zBeeble
The response from CHKDSK is:
"The type of the file system is REFS. The ReFS file system does not need to be checked."

Re: Data corruption error?

Posted: Mon Aug 05, 2019 4:51 pm
by support
Thank you very much for the information! We're checking the compatibility with REFS.