Re: lost of redo logs

Hi Jeff,

I'll confirm that what you report happens on 9i, Windows 2000, so it's not version or O/S specific. However, several cups of tea, an hour or so of early-morning Discovery Channel, and a good book later, the status *did* spontaneously change from 'ACTIVE' to 'INACTIVE' without a second log switch:

SQL> alter system switch logfile;

System altered.

SQL> select group#, archived, status from v$log;

    GROUP# ARC STATUS

---------- --- ----------------
         1 NO  INACTIVE
         2 NO  ACTIVE
         3 NO  CURRENT

[Long, long wait here]

SQL> select group#, archived, status from v$log;

    GROUP# ARC STATUS

---------- --- ----------------
         1 NO  INACTIVE
         2 NO  INACTIVE
         3 NO  CURRENT

Be *extremely* patient, and it will happen. I would imagine, never having investigated it thoroughly, that Oracle is simply playing safe. Since it's 4am here, I'll leave the thorough investigation until later, if that's OK! I'll report back when I have something.

Regards
HJR

--
----------------------------------------------
Resources for Oracle: http://www.hjrdba.com
===============================

[snip]


>

> Why wouldn't a log-switch checkpoint be sufficient to make the current

> logfile inactive?  (Ignore the archiving side of it, as that's happening

> successfully.)

>

> I've got an 8.1.7.3 database on a Solaris box where the previously

"current"


> logfile remains "active" until another logswitch or checkpoint occurs.

> Reading the above and the documentation, I'm led to believe that the

> checkpoint that happens at a log switch should be sufficient to take the

> log from "current" to "inactive" (again, ignoring the archiving side of

it).


> Why then does it take TWO checkpoints to make a logfile inactive?