Page tree
Skip to end of metadata
Go to start of metadata

ASE Traceflag causes ASE to only recover the master database when booting.

This is useful if recovery on some other database is hitting a fatal error that either crashes the server or exits leaving locks on sysdatabases.  By recovering only master, the database causing the problem can be dropped and then the server shutdown and rebooted normally.

Note that most system stored procedures cannot be used because they rely on temporary tables and due to this flag, tempdb has not been recovered,

You will see the following in the errorlog if the server is booted with traceflag 3608:

 

...

00:0006:00000:00000:2014/01/02 09:10:54.62 server  Active traceflags: 3608
00:0006:00000:00001:2014/01/02 09:10:54.63 kernel  libomni1 - Component Integration Services: using 'Sybase Client-Library/16.0/P/DRV.16.0.0.00/SPARC/Solaris 10 Native Threads/BUILD1600-000/64bit/DEBUG/Tue Nov  5 14:48:12 2013'
00:0006:00000:00001:2014/01/02 09:10:54.63 server  Size of the 2K memory pool for the cache (default data cache): 8192 Kb
Memory allocated for the default data cache cachelet 1: 8192 Kb
00:0006:00000:00001:2014/01/02 09:10:54.63 server  Opening Master Database ...
00:0006:00000:00001:2014/01/02 09:10:55.01 server  Loading ASE's default sort order and character set
00:0006:00000:00001:2014/01/02 09:10:55.28 kernel  HADR previous mode before master recovery is : -1
00:0006:00000:00001:2014/01/02 09:10:55.28 server  WARNING: ***************************
00:0006:00000:00001:2014/01/02 09:10:55.28 server  Recovering only master database
00:0006:00000:00001:2014/01/02 09:10:55.28 server  WARNING: ***************************
00:0006:00000:00001:2014/01/02 09:10:55.28 server  Recovering database 'master'.
00:0006:00000:00001:2014/01/02 09:10:55.39 server  Started estimating recovery log boundaries for database 'master'.
00:0006:00000:00001:2014/01/02 09:10:55.47 server  Database 'master', checkpoint=(37023, 12), first=(37023, 12), last=(37023, 12).
00:0006:00000:00001:2014/01/02 09:10:55.47 server  Completed estimating recovery log boundaries for database 'master'.
00:0006:00000:00001:2014/01/02 09:10:55.48 server  Started ANALYSIS pass for database 'master'.
...

 

  • No labels

1 Comment

  1. Anonymous


    Does this mean that my user database which needs to run recovery processes will be offline or in accessible?

     

    is there anyway where we can keep the databases online and at the same time no recovery is done on the database at startup?

     

    Thanks