Re: Help: Standby database issue

From: OracleMS <lostpwds_at_hotmail.com>
Date: 1 Jun 2001 08:13:52 -0700
Message-ID: <17742ed1.0106010713.70a8ff6f@posting.google.com>

eeknoch_at_gci.net (Edward Knoch) wrote in message news:<c242a22d.0105301411.240f7e98_at_posting.google.com>...

> "ff" <cfs3526_at_ureach.com> wrote in message news:<WhKH6.7841$Aj1.3878670_at_typhoon2.ba-dsg.net>...

> > We have a standby databse, and we use a in house developed perl script to
> > copy the redologs from the remote production database.
> > The way we do is we use 'ls -l' to detect the redolog file, as soon as we
> > see it, wait for 30 seconds, and then we copy it over. Most of time it
> > worked, but sometimes, the copy begins before the production finish writing
> > to the redo, and we got corrupted file.
> > Is there any way we can monitor the finish of the redologs and begin to copy
> > only after the writing of redologs have finished?
> >
> > Thanks for your help.
> >
> > PS: Could some body show me your script for manually copy your redo logfile
> > to remote standby database?
> >
> > Thanks.

You can start up a Server Manager/SQL Plus session in your perl script and check one of two views - v$log or v$archived_log (actually, you can also do 'archive log list', but there's unneccessary parsing involved to get the sequence number of the last archived log.)

In v$log, you can check the largest sequence# that has a value of YES in the ARCHIVED column. That is the last archived log. Or in v$archived_log, you can check for the max(sequence) and that will be your last archived log. Since you will have a certain log format, you can reconstruct the log filename in your script (so if your format is sid_sequence#.arc, then by getting the max sequence# from v$archived_log, you can generate the filename as /your/archive/dest/directory/${ORACLE_SID}_001200.arc and then copy all files in your archive destination that have sequence# less than this one over to your backup directory. . You can do a readdir() or glob in Perl to get the filenames in the directory. Put that in a while loop and you can repeat the actions for all the files. This way you dont have to do an 'ls -l'. If you have any problems getting this to work, post the specific nature of the problem and I can give you some code snippets. I would have given one now, but I'm late for work. I'm sure others will post valuable information as well. Best of luck. Received on Fri Jun 01 2001 - 10:13:52 CDT