Dbvisit Community

How does Dbvisit Standby handle network connection timeouts?

Mike Donovan — Thu, 17 May 2012 00:21:41 GMT

Question:
If a network crashes (times out) while the Primary Database server is shipping archivelog files to the Standby Database server what will happen? Do I have to set a timeout in the DDC file? Or is this done in the ssh configuration?

Response:
If, for any reason, there is any network interruption Dbvisit Standby will attempt to re-transfer the logs on its next run automatically, without any intervention on the part of the DBA. Thresholds can be set so that you can specify how many time Dbvisit should re-transfer a log before notifications should be sent.

The key parameter to set in your DDC file is TRANSFER_RETRIES, which we explain in the following:

Determines the number of retries when the checksum of the transferred log file does not match the original file. When the checksum does not match, the file will be transferred again. If the checksum does not match after TRANSFER_RETRIES attempts, an alert will be raised.
Default value: 2
Example:
TRANSFER_RETRIES=2

Furthermore, network timeouts are actually set by the SSH configuration. So Dbvisit Standby relies on the SSH configuration to get the timeout signal back when the transfer was not completed successful. As soon as this is received, then Dbvisit Standby will attempt the transfer again. The important settings as far as Dbvisit Standby is concerned are:
TRANSFER_RETRIES wich was highlighted above.

RETRIES_WAIT_IN_SEC - Wait in seconds before a retry attempt of TRANSFER_RETRIES. By default this is 5 seconds.

ORA-01186, ORA-01122, ORA-01110, ORA-01209 when opening standby as Read Only

Anton — Sun, 06 May 2012 23:47:49 GMT

Recently a site experienced ORA-1209 errors when opening the standby database read-only. A call was logged with Oracle Support and a summary of the solution is provided below:

When you open a standby database as read only, you notice the following in the database alert log.

< ORA-01186: file 5 failed verification tests
< ORA-01122: database file 5 failed verification check
< ORA-01110: data file 5: '/u02/oracle/oradata/prod/data01.dbf'
< ORA-01209: data file is from before the last RESETLOGS
< File 5 not verified due to error ORA-01122
< ORA-16004 signalled during: alter database open read only...

After investigation the following was noted:

File belong to a read-only tablespace called "DATA"
The database was recovered in the past and opened with resetlogs; both primary and standby show the same resetlogs_time in v$database
The datafile checkpoint_time in v$datafile_header show that it is before the v$database.resetlogs_time

To resolve the ORA-01209 to allow you to open the standby database read-only, follow these steps:

On the Primary Database:

Place the DATA tablespace in "Read Write" mode.
Take the datafile 5 offline
Recover datafile 5
Place tablespace DATA back in "Read Only" mode
Switch the current redo log and issue a checkpoint
Run dbvisit standby on both primary and standby

Example Steps on Primary Database:

sql> alter tablespace DATA read write;
sql> alter database datafile 5 offline;
sql> recover datafile 5;
sql> alter tablesapce DATA read only;
sql> alter system archive log current;
sql> alter system checkpoint;

Now run dbvisit on the Primary then Standby database.

You should now be able to open the standby database read only.

Amazon RDS and access to target database

bigjocker — Sat, 05 May 2012 19:00:10 GMT

Hi,

We're evaluating DBVisit Replicate to migrate one of our production databases to Amazon RDS with Oracle (we are DBVisit Standby customers). RDS instances are isolated machines that cannot be accessed via SSH (we can access only via port 1521), so no software can be installed locally.

Is there a way to use DBVisit replicate in this scenario? I've checked the user manual and it requires the installation of the DBVisit Replicate software on both source and target machines, but this is not possible under Amazon RDS.

We can set up a second server on the Amazon cloud, that has access as a client to the RDS instance (full sqlplus access via tnsnames.ora), but the SYS and SYSTEM accounts are a little limited.

Is this a supported environment for DBVisit Replicate?

Thanks

Primary to Standby to Standby

jamie.gadong — Wed, 25 Apr 2012 14:40:32 GMT

Hi to all,

I have a quick question that someone may be able to answer?

I know you can run dbvisit such that for a given primary database you can have more than one physical standby, and I know that the log shipping would go from the primary to each of these standbys, but is it possible for logs to ship to standby 1, then standby 2 be updated from standby 1 rather than the primary, keeping the extra load off the primary server?

Thanks to all who reply.