Just some details on the cleanup issue. I’ve now had a 3rd full image backup run, and the nightly tidy has deleted the second one and not the first??
These are my settings:
This was after two full image backups. On the first night it didn’t remove one, and running cleanup.bat produce the error mentioned in a previous post:
Is this a bug for two reasons?
A) It’s not running tidy every night AND not reducing backups to only ONE as per settings.
B) When it did run a tidy it removed the middle image, but not the oldest one.
It cannot delete a full backup if there is an incremental one based on that and deleting the incremental backups would cause the number of incremental image backups fall below the minimum number of incremental image backups.
Incremental file backup failed on one of our servers with error message “Client calculated hash of “/media/backup/urbackup_btrfs/SV405/160419-2228/isagiwf/Base/T_isagiwf_ZZ.GI_VALDICO.mdf” differs from server calculated hash. This may be caused by a bug or by random bit flips on the client or server hard disk. Failing backup.”, even if we relaunch manually an incremental file backup.
However, if we launch a Full file backup of the client, it goes successfully.
Here’s the full log entry :
19/04/16 22:28
Starting incremental file backup...
Info
19/04/16 22:30
Scanning for changed hard links on volume of "Fichiers"...
Info
19/04/16 22:30
Indexing of "Fichiers" done. 1 filesystem lookups 922 db lookups and 0 db updates
Info
19/04/16 22:30
Indexing of "SavDossiers" done. 1 filesystem lookups 402 db lookups and 0 db updates
Info
19/04/16 22:30
Scanning for changed hard links on volume of "EIC"...
Info
19/04/16 22:30
Indexing of "EIC" done. 1 filesystem lookups 270 db lookups and 0 db updates
Info
19/04/16 22:30
Indexing of "invent" done. 1 filesystem lookups 0 db lookups and 0 db updates
Info
19/04/16 22:30
Indexing of "isagiwf" done. 2 filesystem lookups 2339 db lookups and 1 db updates
Info
19/04/16 22:30
Indexing of "majdos" done. 1 filesystem lookups 1 db lookups and 0 db updates
Info
19/04/16 22:30
Indexing of "majpos" done. 1 filesystem lookups 0 db lookups and 0 db updates
Info
19/04/16 22:30
SV405: Loading file list...
Info
19/04/16 22:30
SV405: Calculating file tree differences...
Info
19/04/16 22:30
SV405: Creating snapshot...
Info
19/04/16 22:30
SV405: Deleting files in snapshot... (25042)
Info
19/04/16 22:30
SV405: Deleting files in hash snapshot...
Info
19/04/16 22:30
SV405: Calculating tree difference size...
Info
19/04/16 22:30
SV405: Linking unchanged and loading new files...
Info
19/04/16 22:32
Waiting for file transfers...
Errors
19/04/16 22:34
Client calculated hash of "/media/backup/urbackup_btrfs/SV405/160419-2228/isagiwf/Base/T_isagiwf_ZZ.GI_VALDICO.mdf" differs from server calculated hash. This may be caused by a bug or by random bit flips on the client or server hard disk. Failing backup.
Info
19/04/16 22:34
Waiting for file hashing and copying threads...
Info
19/04/16 22:34
Writing new file list...
Info
19/04/16 22:34
Number of readded file entries is 182048
Errors
19/04/16 22:34
Fatal error during backup. Backup not completed
Info
19/04/16 22:34
Transferred 42.7303 MB - Average speed: 1.41118 MBit/s
Info
19/04/16 22:34
(Before compression: 404.047 MB ratio: 9.45573)
Info
19/04/16 22:34
2.55278 GB of files were already present on the server and did not need to be transferred
Errors
19/04/16 22:34
FATAL: Backup failed because of disk problems
Info
19/04/16 22:34
Time taken for backing up client SV405: 6m 22s
Errors
19/04/16 22:34
Backup failed.
What should be the matter ? How can we fix this ? Storage for backup server is BTRFS, server 2.0.10, client Win 2012R2 64 bit with client 2.0.9 installed.
ERROR: Client calculated hash of “some file” differs from server calculated hash. This may be caused by a bug or by random bit flips on the client or server hard disk. Failing backup.
and
ERROR: Fatal error during backup. Backup not completed
ERROR: FATAL: Backup failed because of disk problems
while doing incremental file backups of a Windows 7 computer. Running a full file backup seems to fix it.
Meanwhile, the full & incremental image backups are fine, so I don’t think the hard drive is bad…
Setup: Server 2.0.10 w/BTRFS & Client Windows 7 Pro x64 2.0.9.
Maybe fixed in server 2.0.11 / client 2.0.10?
Edit:
I forgot to mention: once this happens with a particular file, it keeps happening with the same file for each incremental file backup.
The server is also on a Xeon CPU / ECC RAM / HW RAID controller w/protected cache, so the probably of a bit flip on the server is probably fairly low…The workstation however, anything’s possible…
Yes this is the same for me, keeps hapenning on the same for every incremental file backup. Both server and client runs as VMs on a Prxmox KVM cluster. No problems with others VMs from other offices with similar setup / hardware …
@Greg@TomTomGo Is this a new server installation or an upgrade from 1.4.x or an early beta server version?
At this point the bug could be in a lot of components. Assuming it is on the server side it would help if you send me the different file versions if you can.
I’d need the file from the last complete file backup (full or incremental) before it started failing. Plus the hash file at the same path in the .hashes folder. And the file it tried to transfer, but failed. If you ran a full backup, it should be in the full backup.
The server was originally fresh installed with 2.0.7 beta and 2.0.8 beta for the client. How can i send you the files because they are about 4 Gb each (it’s a SQL Server mdb …).
