Re: Interpreting some alrt.log output

Ed Stevens <nospam_at_noway.nohow> wrote in message news:<f40r70582idcsc88bl6p61ngfa34d70so8_at_4ax.com>...
> Platform: Oracle 8.1.7 EE on Solaris 8
>
> No particular problem, just poking around in the alert log to get a
> better understanding of exactly what normal output is telling me.
> Here is a chunk of one. My comments/observations/questions are
> enclosed in parenthesis. Apologies for any line wrap. Would anyone
> care to set me straight if my obervations are wrong? Thanks.
>
> Tue Apr 13 20:11:57 2004
> Beginning log switch checkpoint up to RBA [0x200.2.10], SCN:
> 0x0000.01957b81
> Thread 1 advanced to log sequence 512
> Current log# 2 seq# 512 mem# 0: /u06/oradata/CMQM1240/redo201.log
> Current log# 2 seq# 512 mem# 1: /u07/oradata/CMQM1240/redo202.log
>
> (we just switched to log group #2. This is the 512th log switch since
> the db was put in archive log mode. ??)

Well, look at the sequence number just after the CREATE DATABASE statement (assuming you created it in archive log mode). My first switch said 2, presumably 1 was creating the system ts. So the number of switches would be one less than the sequence number?

>
> Tue Apr 13 20:11:57 2004
> ARCH: Evaluating archive log 1 thread 1 sequence 511
> ARCH: Beginning to archive log 1 thread 1 sequence 511
>
> (Since we are now using rdo log group #2 for seq. 512, we can start
> archiving sequence 511, presumably from rdo log group #1
>
> Tue Apr 13 20:11:57 2004
> ARC0: Evaluating archive log 1 thread 1 sequence 511
> ARC0: Unable to archive log 1 thread 1 sequence 511
> Log actively being archived by another process
>
> (A second archive of seq. 511? Why?)

Archiver starts up other processes for various reasons, such as one is taking too long for the load.

>
> Tue Apr 13 20:11:57 2004
> Creating archive destination LOG_ARCHIVE_DEST_1:
> '/u08/arch/CMQM1240/arch_1_511.arc'
> ARCH: Completed archiving log 1 thread 1 sequence 511
>
> (Created and completed writing the archive log for seq. 511)
>
> Tue Apr 13 20:26:45 2004
> ARCH: Possible network disconnect with primary database
>
> ( huh??)

All sorts of things can go wrong, may or may not mean anything. Do you have standby, or multiple arc destinations? One must always assume a network will munge things up, and 8i standby architecture doesn't quite understand this. Worst case, it can hang your primary db, so definitely open a tar and get after your network admins if this happens repeatedly.

>
> Tue Apr 13 20:27:15 2004
> Completed checkpoint up to RBA [0x200.2.10], SCN: 0x0000.01957b81
> Wed Apr 14 08:46:42 2004
> Beginning log switch checkpoint up to RBA [0x201.2.10], SCN:
> 0x0000.01969b1d
>
> (checkpoint and log switch, but notice the time elapsed between the
> two)

That's why I like to use cron to harden the times between log switches with an explicit alter system switch logfile.

>
> Thread 1 advanced to log sequence 513
> Current log# 3 seq# 513 mem# 0: /u06/oradata/CMQM1240/redo301.log
> Current log# 3 seq# 513 mem# 1: /u07/oradata/CMQM1240/redo302.log
> Wed Apr 14 08:46:42 2004
> ARC0: Evaluating archive log 2 thread 1 sequence 512
> ARC0: Beginning to archive log 2 thread 1 sequence 512
> Creating archive destination LOG_ARCHIVE_DEST_1:
> '/u08/arch/CMQM1240/arch_1_512.arc'
> ARC0: Completed archiving log 2 thread 1 sequence 512
>
> (this time there was no second attempt to archive)
>
> Wed Apr 14 09:02:10 2004
> Completed checkpoint up to RBA [0x201.2.10], SCN: 0x0000.01969b1d

How big is that SGA, anyways?

jg

--
@home.com is bogus.
I'll never look at logical volume managers the same again: 
http://ideas.4brad.com/archives/000064.html