get_trc.sh – a easy packing tracefile script
No Comments

Posted by Milo on July 4, 2014 | Short Link

As many of you may know, sometimes when you’re doing a health check on Oracle database. Trace files are needed for analyzing complicated errors.

Are you having trouble on fetching trace file again and again when you go through the error messages one by one OR you gather the Database information from client side, sometimes, you may find you’re not getting the tracefiles, but hard to ask customers to do this job for you.

 

Good news is that we have get_trc.sh script for doing this for us only need a simple configuration, it’s totally help you gather important information, all you need is a FTP SERVER on your computer.

 

Here is the script with v0.6 version till 2014.07.04:

 

The only part you may need to modify, it’s this part:

CURRENT_YEAR : the year where start from alert log.

BEGIN_MONTH  : month in alert log, which wanna make it  a start point.

Here, I am not gonna use day, because in some idle db, we might get the day right, and waste time to guess.

FTP_SER,FTP_USER, FTP_PASSWD :  related with the FTP access, the ip address of FTP SERVER, the upload user name and password.

TARGET_DIR :  is where you need to put the gathering stuffs on the FTP SERVER.

 

You can use this script as this way:

e.g.

The log will show like this:

 

It’s cool so far, right? :)

NOTE: there might be a issue on extend regular expression with grep (Not support ‘-E’ option on solaris 10 ) or egrep(it seems the same syntax works for linux, AIX, but not on solaris 10), it’s a little bit weird. I will keep trying to fix this issue.

 

SCRIPT TESTED PASS ON PLATFORM: AIX 5.3, RHEL (OEL) 5,6.

Please TAKE YOUR RISK TO USE THIS SCRIPT!

If you have any feedback, please give me a comment or send me an email.

Thanks and Good luck.

 

Tags:

CRS-4013, CRS-4000 occurred when stop and start cluster
No Comments

Posted by Milo on June 19, 2014 | Short Link

A very interesting case :

Oracle 11.2.0.3 RAC on HPUX IA64. When I ran DBCA to create a new database on existing environment. This error come out:

This is very strange error on a RAC, so I decide to restart the crs on each node. The crs_stat -t result on node 1 looked cool, but on node 2:


NOTE: There is different on this environment, host test1 has instance testdb2, while testdb1 is on host test2.

 

Stop the cluster error occured:


Oh, it sucks! I also looked into the css and crs log, no any log added.

When I reboot the machines, still no any output from log files.

After googling a lot, Yogesh Bhandarkar gave a workaround for this issue. Thanks  Yogesh. :)

 

Now let’s see what we gotta do.

Let’s deconfig the clusterware:


 

Now let’s rerun the root.sh script:

 

Now after the root.sh complete, clusterware goes to normal status.


I haven’t been figure out what’s real going on, but anyway, this is the way we fixed this.

 

 

Tags: , , , ,

A Procedure transalate hex or dec to ascii string
No Comments

Posted by Milo on June 12, 2014 | Short Link

 

Test this with the following code:

 

At last, the procedure will show you the final string you want to translate, but remember if you’re passing the decimal ascii code, then you should add the second parameter with 10.

Good luck.

 

 

 

 

 

Copyright © 2014-2017 Come on, Milo! All rights reserved.
This site is using the Shades theme, v2.2.1.1, from BuyNowShop.com.