FAILED MESSAGE

IF recover database has been enabled for that thread, then your message should be in the error database and you should be able to find out why the message failed and should be able to resend the message to a file there.  Most the time I just fix the problem and then resend the message from the original inbound thread in SMAT.

You can delete the message from the Error Database using the Database Administrator tool.   Just select the Delete messages option midway down the page in the tool.  As far as resending the message, I usually just look up the message in SMAT by finding the MSG ID and then resend.

Gerand,

If you type in just hcidbdump at the command line, it will show you all the options for the command.  They are all very helpful.

If you want to look at the error database then type in “hcidbdump -e” at the command line.  This will show you all messages that have failed or errored out.

If you want to look at a particular message that is in the error database then you can do the following:

hcidbdump -e -m 0.0.12345 -L

This will give show you the message plus a very brief description of why the message failed (i.e. Tcl callout error).  The -m is the message ID number for the message you to view.   If you want more information as to why the message failed, you will need to look into the process log file.  

Let me know if you need any more info.

Tom Rioux (yep, another Louisiana boy!)

Gerand, just a couple cautions about working with recovery or error db records.  These seem to be the case at least with our environment (AIX 5.2, Cloverleaf 5.4), and possibly whatever version you have.

The gui is fine for working with records as long as there aren’t a huge number of records, or the records are very large and you have selected to display all the details.  In those situations we have seen the gui get locked up (some buffer gets maxed out?), and afterwards we have a dbvista error condition that has to be cleaned up (which is disruptive in a production site).

The same issue sort of applies when running the hcidbdump command.  It is important that you don’t interrupt the command (like ctrl-c) while it is executing, or the next time you try to access the db you will find you have a dbvista error.  It was recommended to us that we always pipe the output of the command to ‘less’, and then use that to page through the output after the command has completed:

hcidbdump -r | less

One thing the gui is great for is to select the various options and filtering you want on the screen, then rather than execute the screen and risk it locking up, instead click the “show command” button, and then copy/paste it to the command line and add on a pipe of the output to less.

Gerand,

We almost always use hcidbdump to look and work with these.  Here is some more info.

1) hcidbdump -e  writes the messages to a file

2) hcidbdump -e -D deletes the messages in the db

3) if you use the -m option, you do not need the 0.0.  all you need is just the number after the second .

4) NEVER NEVER ctrl-C to end it.  We have locked down the entire site by doing this.  No messages could get through any where..

Good Luck.