Db_Vista database error -24: ‘PROGRAMMER/USER error: -24

[Anonymous]

Hi Clovertechies,

A new error message I encountered a couple of days ago.

Each day different processes are panicking. Some of them are not in inter-process communications with multiple processes.

Has anyone seen this error?

How do you remedy this?

TIA

Reggie

Topic

[Scherie Drewa]

I got the exact same error today, and all processes went down.

What does this error mean?

[Anonymous]

Do you have inerprocess communiation setup.

That means data feeds from processA -> processB.

No one has replied to my posting yet!

[Scherie Drewa]

Yes, we do, and have had for a long time.  i am on 3.8.1, and have never gotten this error before.  I am amazed that no one has responded to you, although I hadly use this posting site.

[Anonymous]

Are there any state 14 messages in your database ?

[Scherie Drewa]

No state 14 messages at this time, but there could have been this morning when this happened.

[Tom Patton]

I just upgraded to AIX 5.2 on QDX 5.3 rev3 and just started getting the same error.  Timing on one thread is consistent with an known backup time when the conenction is interrupted, but the prod side for this connection stayed up.

No state 14 msgs, but many state 11.

I should also indicate that I failed over to a server that has never run this version on QDX on this server.  I’ve been running on the primary server only and just swung the SAN over to the secondary Sunday morn.

A lot of variables for me….

[Anonymous]

Tom,

did you look in the .err file. Also check in the

$HCISITEDIR/exec/errors

$HCISITEDIR/exec/hcimonitord/hcimonitord.err file

Does any of the err files shows broken pipe err? If so there is a failure in the interprocess communication.

In my case, I have always had routing from thread_A of process A to thread_B of process B.

Check the nmon/top as well. How does you kernel parameters look.

I am wondering this has something to do with the number of openfiles.

But could not pinpoint to that yet.

Please post your findings.

thanks

Reggie

[Tom Patton]

Reggie,

Thank you for the post.

I did check the process logs and err files.

I have broken pipe errors and yes, these are processes that have inbound connections from another site/process.

I didn’t have any files in  $HCISITEDIR/exec/errors and my

$HCISITEDIR/exec/hcimonitord/hcimonitord.err file didn’t show any unusual errors.  Just “normal errors”  e.g:

06/12/2006 11:13:57 [cmd :cmd :WARN/0:  hcimonitord] alerts client 0x20b98678

06/12/2006 11:13:57 [aler:aler:WARN/0:  hcimonitord] Creating AlertAction: eocno

tify

06/12/2006 11:13:57 [cmd :cmd :WARN/0:  hcimonitord] eonotify client 0x20b98678

My limits fie has the following which should be ok for # of open files:

default:

      fsize = 2097151

      core = 2097151

      cpu = -1

      data = 1572864

      rss = 393216

      stack = 393216

      nofiles = 10000

[Anonymous]

Tom,

Your Kernal Parameter values are same as ours.

Broken Pipe error shows that the interprocess communication failed.

Next time when the panic / DB_Vista -24 error occurs, what you may want to do before starting the processes, use lsof command to see whether there are any open files.

One last question, are you routing from Thread A of Process A of SITE1 to another SITE ?

Please let me know what you find.

thanks

Reggie

[Tom Patton]

Thanks Reggie, I’ll post lsof info if I receive another error.

And yes, I have A,B and C threads in one site that feed A,B and C threads in another site.

[Marv Beck]

Hi Tom:

I didn’t see any further postings for your question.  I was wondering if you have a resolution.  I’m getting set to do a “loop back” interface sometime this week and your issue could boil up on my interface engine.  However, I’m not sending to another site.  I’m going to be sending outbound from process A to an inbound of process B so I can avoid a cross process connection

TIA,

Marv

[Author]

Replies