| Standby database [message #61773] |
Wed, 26 May 2004 08:37  |
Bill Wallace
Messages: 1
Registered: May 2004
|
Junior Member |
|
|
We are running Oracle 8.1.7 on Solaris 7 and are planning to install a standby database on a remote host...using 'Managed Recovery Mode' ( ie; using LOG_ARCHIVE_DEST_3 = 'SERVICE=stby').
There will be periods of increased activity on the primary database when the tcp link to the standby will not be able to keep up with the frequency of log switches.
We are concerned that transmission of log 'n' to the standby will not be completed before the switch of log 'n+1' occurs on the primary.
Will this situation cause the primay database to hang?
|
|
|
|
| Re: Standby database [message #61775 is a reply to message #61773] |
Wed, 26 May 2004 12:27  |
ctg
Messages: 146
Registered: July 2002
|
Senior Member |
|
|
standard dba answer: IT DEPENDS.
it could lead the database to hang. it will depends on how many groups, how long does it take to fill the members, how long does it take to copy the members over.
Scenario 1:
you have 2 redo log groups
switch from group 1 to group 2
group 1 member sent to standby
switch from group 2 to group 1
if group 1 has not finished being shipped to standby then your db will hang.
Scenario 2:
you have 3 redo log groups
switch from group 1 to group 2
group 1 member sent to standby
switch from group 2 to group 3
if group 1 has not finished being shipped to standby then your db will NOT hang because it does not need group 1 yet
switch from group 3 to group 1
if group 1 has finished being shipped to standby you are ok. but if it still hasnt finished, then your db will hang.
HTH
|
|
|
|
| Re: Standby database [message #61780 is a reply to message #61773] |
Wed, 26 May 2004 22:19 |
ora9iDBA
Messages: 31
Registered: March 2003
|
Member |
|
|
Upgrade your database to 9.2.0.5 and in dataguard,use
_LOG_ARCHIVE_CALLOUT='LOCAL_FIRST=TRUE' parameter. It will resove your problem
Kishor
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
