View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000451 | bareos-core | General | public | 2015-04-03 14:16 | 2015-07-26 12:05 |
Reporter | SJ49 | Assigned To | |||
Priority | normal | Severity | major | Reproducibility | always |
Status | closed | Resolution | unable to reproduce | ||
Platform | Linux | OS | CentOS | OS Version | 7 |
Product Version | 14.2.2 | ||||
Summary | 0000451: Can't restore NDMP backup | ||||
Description | Hello, We have a Netapp controller we want to backup to tape with NDMP protocol. Actually, we can backup but never restore with ndmp. That's ok with classic mode. The error message is on restore : Fatal error: Read storage rbt01-ndmp doesn't point to storage definition with paired storage option. We have try with Backup Format = dump and Backup Format = smtape and it's the same error. Mains configuration parts in Additional Information . Thanks, SJ | ||||
Additional Information | ### bareos-sd.conf : Storage { # definition of myself Name = xbareos2-sd Maximum Concurrent Jobs = 1 NDMP Enable = yes NDMP Port = 10001 } Director { Name = xbareos2-dir Password = "SQvRcOo2FhTMcDvAqpEqNXesFvKMHvXiOMyQP8cp4emm" } Director { Name = xbareos2-mon Password = "1GYwDXUi+7f9cbmFdAwOLss9wARjet0aNa1Ng5+G3wCF" Monitor = yes } Messages { Name = Standard director = xbareos2-dir = all } Ndmp { Name = xbareos2-ndmp-dma Username = root Password = cg49cg49 AuthType = Clear } Device { Name = tapedrive0 # DeviceType = tape Drive Index = 0 Archive Device = /dev/st0 Media Type = LTO-5 AutoChanger = yes # default: no AutomaticMount = yes # default: no LabelMedia = yes RemovableMedia = yes AlwaysOpen = yes Random Access = No } Autochanger { Name = autochanger0 Changer Device = /dev/sg3 Device = tapedrive0 Changer Command = "/usr/lib/bareos/scripts/mtx-changer %c %o %S %a %d" } | ||||
Tags | No tags attached. | ||||
I get the feeling you are doing an override of the storage on restore. How things normally work is that on backup of NDMP things are saved to the "native" storage of the SD e.g. the one paired with the NDMP storage. On restore it will see things are stored on a native storage device and look for the paired NDMP device that belongs to that storage. You probably need the "rbt01" storage as read storage. |
|
Hello, Thanks for your reply. I try to use the read storage "rbt01" as you indicate in you post for the restore but that doesn't work. I receive the following error : ****************************************************************** 13-Apr 09:29 xbareos2-dir JobId 49: Start Restore Job NDMPRestoreDump.2015-04-13_09.29.37_07 13-Apr 09:29 xbareos2-dir JobId 49: Using Device "tapedrive0" to read. 13-Apr 09:29 xbareos2-dir JobId 49: Opening tape drive OLME-KHPA-MIHM-OMEB-DGLB-CBIF-BGMK-GFJC@/vol/vol_testbareosCIFS read-only 13-Apr 09:29 xbareos2-dir JobId 49: Commanding tape drive to rewind 13-Apr 09:29 xbareos2-sd JobId 49: Ready to read from volume "E02003L4" on device "tapedrive0" (/dev/st0). 13-Apr 09:29 xbareos2-sd JobId 49: Forward spacing Volume "E02003L4" to file:block 1:0. 13-Apr 09:29 xbareos2-dir JobId 49: Waiting for operation to start 13-Apr 09:29 xbareos2-dir JobId 49: Async request NDMP4_LOG_MESSAGE 13-Apr 09:29 xbareos2-dir JobId 49: Operation started 13-Apr 09:29 xbareos2-dir JobId 49: Monitoring recover 13-Apr 09:29 xbareos2-dir JobId 49: DATA: bytes 0KB MOVER: read 0KB record 0 13-Apr 09:29 xbareos2-dir JobId 49: DATA: bytes 0KB MOVER: read 0KB record 0 13-Apr 09:29 xbareos2-dir JobId 49: Mover paused, reason=NDMP9_MOVER_PAUSE_EOF 13-Apr 09:29 xbareos2-dir JobId 49: End of tapes 13-Apr 09:29 xbareos2-dir JobId 49: LOG_MESSAGE: 'RESTORE: Read Socket received EOF.' 13-Apr 09:29 xbareos2-dir JobId 49: LOG_MESSAGE: 'RESTORE: RESTORE IS ABORTED' 13-Apr 09:29 xbareos2-dir JobId 49: DATA: bytes 0KB MOVER: read 0KB record 0 13-Apr 09:29 xbareos2-dir JobId 49: Error: LOG_MESSAGE: 'DATA: Restore terminated: EVENT: I/O ERROR (for vol_testbareosCIFS)' 13-Apr 09:29 xbareos2-dir JobId 49: DATA: bytes 0KB MOVER: read 0KB record 0 13-Apr 09:29 xbareos2-dir JobId 49: Operation done, cleaning up 13-Apr 09:29 xbareos2-dir JobId 49: Waiting for operation to halt 13-Apr 09:29 xbareos2-dir JobId 49: Commanding tape drive to rewind 13-Apr 09:29 xbareos2-dir JobId 49: Closing tape drive OLME-KHPA-MIHM-OMEB-DGLB-CBIF-BGMK-GFJC@/vol/vol_testbareosCIFS 13-Apr 09:29 xbareos2-dir JobId 49: Operation halted, stopping 13-Apr 09:29 xbareos2-dir JobId 49: Operation ended questionably 13-Apr 09:29 xbareos2-dir JobId 49: Error: Bareos xbareos2-dir 14.2.2 (12Dec14): Build OS: x86_64-redhat-linux-gnu redhat CentOS Linux release 7.0.1406 (Core) JobId: 49 Job: NDMPRestoreDump.2015-04-13_09.29.37_07 Restore Client: netapptf-2-ndmp Start time: 13-Apr-2015 09:29:39 End time: 13-Apr-2015 09:29:42 Elapsed time: 3 secs Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s SD termination status: Running Termination: *** Restore Error *** ****************************************************************** I don't know what's wrong. Once again, normal restores work (for exemple restore of the catalog). SJ |
|
Ok seems things work fine now as you now run into: https://bugs.bareos.org/view.php?id=374 which is fixed in 14.2.3 and 15.1.0 (experimental). So either compile 14.2.3 or use 15.1.0 or buy a subscription and get 14.2.4. |
|
Thank you for your reply, I update to bareos 15.1.0 but now I can't backup anymore NDMP. I got the following error : 14-Apr 16:25 xbareos2-dir JobId 11: Commanding tape drive to rewind 14-Apr 16:25 xbareos2-dir JobId 11: ERR-CONN NDMP4_MOVER_LISTEN exchange-failed 14-Apr 16:25 xbareos2-dir JobId 11: Error: Director's comm line to SD dropped. I keep the same configuration files, is-there any change involved by v15.1 ? Once again, backups of the catalog (ie standard backup) are still ok. |
|
15.1 has some extra fixes and checks for blocksize that are needed for SMTAPE as NetAPP allows you to set quite some things but then fail to restore so the blocksize needs to be in multiples of 4 Kb to a maximum of 240 Kb. But the error above seems to be some NDMP handshake error where the DATA MOVER doesn't want to do something (could be related to blocksize) but a NDMP debug session will probably show your actual problem. |
|
Date Modified | Username | Field | Change |
---|---|---|---|
2015-04-03 14:16 | SJ49 | New Issue | |
2015-04-03 14:16 | SJ49 | File Added: bareos-dir.conf | |
2015-04-11 10:06 | mvwieringen | Note Added: 0001683 | |
2015-04-11 10:07 | mvwieringen | Status | new => feedback |
2015-04-13 09:37 | SJ49 | Note Added: 0001684 | |
2015-04-13 09:37 | SJ49 | Status | feedback => new |
2015-04-13 10:50 | mvwieringen | Note Added: 0001685 | |
2015-04-13 10:50 | mvwieringen | Status | new => feedback |
2015-04-14 16:35 | SJ49 | Note Added: 0001686 | |
2015-04-14 16:35 | SJ49 | Status | feedback => new |
2015-04-16 22:44 | mvwieringen | Note Added: 0001690 | |
2015-04-16 22:44 | mvwieringen | Status | new => feedback |
2015-07-26 12:05 | mvwieringen | Status | feedback => closed |
2015-07-26 12:05 | mvwieringen | Resolution | open => unable to reproduce |