on 07-25-2013 12:41 PM
Hi all ,
I excuted Hana point-in-time recovery with HANA-Studio .
But the recovery could not be completed . And there was the error message as below .
"The service 'nameserver' at 'tyoscXXX:30001' responsible for the volume '1' does crash "
And I checked the backup log . Below are the error message's content.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY from file: log_backup_2_0_115990656_115996096 in /usr/HANAlog/
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY from file: log_backup_2_0_115996096_116001472 in /usr/HANAlog/
・
・
・
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY from file: log_backup_4_0_3375936_3377024 in /usr/HANAlog/
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY from file: log_backup_4_0_3377024_3378112 in /usr/HANAlog/
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY from file: log_backup_4_0_3378112_3379200 in /usr/HANAlog/
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY state of service: nameserver, tyoscXXX:30001, volume: 1, RecoveryExecuteTopologyRecoveryInProgress
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY state of service: nameserver, tyoscXXX:30001, volume: 1, RecoveryExecuteTopologyRecoveryFinished
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY start of progress monitoring, volumes: 4, bytes: 0
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY state of service: nameserver, tyoscXXX:30001, volume: 1, RecoveryExecuteDataRecoveryInProgress
2013-07-25T19:26:55+09:00 P11273 140155dfa1d INFO RECOVERY progress of service: nameserver, tyoscXXX:30001, volume: 1, 0% 0/89735168
2013-07-25T19:26:57+09:00 P11273 140155dfa1d INFO RECOVERY progress of service: nameserver, tyoscXXX:30001, volume: 1, 99% 89731072/89735168
2013-07-25T19:26:59+09:00 P11273 140155dfa1d INFO RECOVERY progress of service: nameserver, tyoscXXX:30001, volume: 1, 100% 89735168/89735168
2013-07-25T19:26:59+09:00 P11273 140155dfa1d INFO RECOVERY state of service: nameserver, tyoscXXX:30001, volume: 1, RecoveryExecuteDataRecoveryFinished
2013-07-25T19:26:59+09:00 P11273 140155dfa1d INFO RECOVERY state of service: nameserver, tyoscXXX:30001, volume: 1, RecoveryLogReplayStarted
2013-07-25T19:27:05+09:00 P11506 140155dfa1d ERROR RECOVERY RECOVER DATA finished with error: [448] recovery could not be completed, volume 1, reached log position 0, [110044] The service 'nameserver' at 'tyoscXXX:30001' responsible for the volume '1' does crash
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
What was occured ?
Please give me some advice .
Thanks in advance .
Best Regards .
Kazuki
Hi Kazuki,
You can check the following Note:
Note 1642148 - FAQ: SAP HANA Database Backup & Recovery
At the end of this note, you can find other notes related to Recovery getting failed
Regars,
Vivek
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Vivek,
Thank you for reply . I checked Note 1642148 and I got Note 1779221 - Recovery of SAP HANA database fails .
I have one question . When I excuted Hana point-in-time recovery with HANA-Studio , I was instructed to specify location of log backup files to be used to recover the database. So I specified network drive that transferd Log backup from SSD .Was this mistake ? Should I specify SSD ? But SSD has 4 area . Which area should I specify ?
Thanks in advance .
Best Regards .
Kazuki
Hi Kazuki,
Please refer note 1779221.
Reason and Prerequisites
The failed recovery should restore the database to the most recent state or to a specific point in time.
Furthermore, in the history of the database, there was already at least one recovery with a selected data backup without log backups being implemented (recover ... to a specific data backup). After this recovery, the database started a new log sequence whose log items overlapped with an existing log.
As of revision 46, this problem should be solved by selecting the required log backup sequence in the SAP HANA Studio. Otherwise, you must proceed as described below.The log backups of Revision 25 and older do not have properties that allow the selection of the required log backup
sequence in the SAP HANA Studio. You must proceed as described below in this case also.
Solution
The selection of the required log backup sequence must be made manually. For this, unwanted log backup files must be moved in the file system, the backup catalog must be regenerated, and the recovery must be repeated.
1. Moving the unwanted log backups
You can use only a log that was generated either before the first recovery described above or that was generated after the recovery. Each unwanted log backup file must be moved.
Refer to the file backup.log for the point in time of the recovery. There the recovery command "RECOVER DATA USING FILE ('<backupprefix>') CLEAR LOG" is logged with a time stamp.
The point in time of the recovery helps you to locate the interruption of the log backup sequence.
If there were several recoveries that interrupted the log, then there are more incompatible sequences in the log accordingly. For two recoveries, there is one sequence that occurred before the first recovery; one sequence between the first and the second recovery and one sequence after the second recovery.
Alternatively, the switch of log sequences can be determined in the file system. The log backup files contain the log items in the file name and the sequence of the log items was interrupted by the recovery. The log backup files are named according to the model log_backup_<volume>_0_<pos1>_<pos2>. The files of a volume produce a sequence in which <pos2> of a file occurs as <pos1> of a successor file. If the log sequence was interrupted, this sequence is also interrupted.
The log backups are in a directory that is configured by the parameter basepath_logbackup.
Example (only log backups of the name server):
> ls -ltr log_backup_1*
-rw------- 1 r25adm sapsys 20480 2012-10-23 15:35 log_backup_1_0_381056_381248
-rw------- 1 r25adm sapsys 221184 2012-10-23 15:36 log_backup_1_0_381184_384512
-rw------- 1 r25adm sapsys 12288 2012-10-23 15:36 log_backup_1_0_384512_384576
-rw------- 1 r25adm sapsys 45056 2012-10-23 15:41 log_backup_1_0_384576_385152
-rw------- 1 r25adm sapsys 1744896 2012-10-23 16:00 log_backup_1_0_385152_412288
-rw------- 1 r25adm sapsys 376832 2012-10-23 16:17 log_backup_1_0_412288_418048
In the example, the log was interrupted at 15:36. The file log_backup_1_0_381056_381248
does not have a successor. The example shows only the log backups of the name server or of volume 1. The log backups of all services or volumes must be taken into account. The example shows only one interruption of the log.
Decide which log backup sequence you want to restore and move the unwanted files to
another directory. This directory must not be a subdirectory to a directory that contains the required log backup files.
2. Generating the backup catalog
As of Version SP 5, SAP HANA database uses an existing backup catalog for the
recovery. If you use Version SP 5 of SAP HANA database or higher, you must
regenerate the backup catalog on the basis of the required backups.
Skip this step if you use SP 4 or lower.
Use the program hdbbackupdiag to regenerate the backup catalog. Use the following options:
-d : Value of the parameter basepath_logbackup as the directory to which the backup catalog is to be saved
-c : log_backup_0_0_0_0.00 as the name for the backup catalog
--dataDir : Directory in which the required data backup is located
--logDirs : List of directories (separated by commas or spaces) in which the
required log backups are located
Example:
hdbbackupdiag -d $DIR_INSTANCE/backup/log -c log_backup_0_0_0_0.00 --generate --dataDir $DIR_INSTANCE/backup/data --logDirs $DIR_INSTANCE/backup/log
3. Repeating the recovery
Now repeat the recovery. If you have chosen the newest log backup sequence in step 1, you can freely select the option "initialize log area" for the recovery. Otherwise, the option must always be selected because the log area of the database is not compatible with the selected log backups sequence.
Hope this helps.
Regards,
Kalyan.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Kalyan,
Thank you for reply .
I checked note 1779221 and excuted the below procedure .
First , I checked whether there are the interrupted log . There are not the interrupted log .
Second , I regenerate the backup catalog with the program hdbbackupdiag .
I removed old backup catalog and genarated new backup catalog in Location of log backup files .
(Because there are old backup catalog named log_backup_0_0_0_0.00~log_backup_0_0_0_0.09 , I generated new backup catalog named log_backup_0_0_0_0.00~log_backup_0_0_0_0.09 .)
Third , I repeatd the recovery .
(Without the option "initialize log area" .)
But the same error was still occured.
Are there any points that are wrong in the above procedure ?
Thanks in advance .
Best Regards .
Kazuki
HiOhya,
Please use command line
Recover the database until a point in time with a timestamp using a dedicated directory for
data backups and further directories containing automatically written log backups.
Use the following command:
RECOVER DATABASE UNTIL TIMESTAMP '2011-08-22 15:00:00' CHANGE ALL DATA USING PATH
('/backup/MONDAY/') CHANGE ALL LOG USING PATH ('/backup/logs1/','/backup/logs2/')
BR
Atul
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
85 | |
10 | |
10 | |
9 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.