on 08-12-2015 4:18 PM
Hi Experts,
Could you make me understand system point in restore using Sybase trans backup.
I have scheduled on-line backups at 12 AM everyday and Trans backups at every 1 hour. Suppose I need to restore system to 1:30 AM is it possible to restore to that particular point in time? Because I will be having the Trans backup at 1 AM.
How can I bridge the gap from 1 AM TO 1:30 AM. Could you please explain me the possibilities.
Best Regards
Amarnath.
Assuming you had made it past 2AM and had a transaction log dump taken at that time, what you would do is:
load database <dbname> from <12AMFullDatabaseDump>
load transaction <dbname> from <1AMTransactionDump>
load transaction <dbname> from <2AMTransactionDump> with until_time = "1:30 AM <date>"
If you had taken a full database dump as 12AM and a transaction log dump at 1AM and then experienced a failure of a data device before the 2AM transaction log was taken, you would do
dump transaction <dbname> to <timeOfFailureTransactionDumpFile> with no_truncate
go
then:
drop database <dbname>
create database <dbname> on <device_specifications> for load
load database <dbname> from <12AMFullDatabaseDump>
load transaction <dbname> from <1AMTransactionDump>
load transaction <dbname> from <timeOfFailureTransactionDumpFile>
go
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Bret,
Thanks a lot for the apt solution.
Could you please help me with my understanding on this particular command
"dump transaction <dbname> to <timeOfFailureTransactionDumpFile> with no_truncate"
What exactly <time of failure transaction dump file mean> ? Is that referring to transaction backup that has to happen on 2 AM?
Please make me understand.
Best Regards,
Amarnath.
It is just the arbitrary file name you would use to take one last transaction log dump in a situation where you have experienced a data device failure. The "with no_truncate" option allows the tran dump to happen even though the database is now in a corrupt state.
This isn't a very commonly encountered scenario - it is far more common for people to use the "with until_time" method to recover the database to the state it was in just before someone did something bad like unexpectedly drop an important table.
User | Count |
---|---|
88 | |
10 | |
10 | |
9 | |
7 | |
7 | |
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.