SYSTEM ERROR: latch 14 depth -1. (3715)

lenindasan

New Member
Hi,
My database goes to ABNORMAL shutdown with the following error.

" SYSTEM ERROR: latch 14 depth -1. (3715) "

my server configuration details are as follows

Platform: ia64 hp server rx3600
OS: B.11.23
Progress : 91D09

Kindly send your suggestion to prevent from ABNORMAL DB shutdown

Thanking you in Advance
 
Stop using kill -9.

And upgrade to a current release of Progress. If you insist on staying with version 9 then at least get upgraded to 9.1e04 which is several years old now and the very last v9 release & service pack -- there will be no more.
 
Thank you for your reply Mr.TOM.

1.We are not using 'KILL -9' .

But still very frequently the DB goes for ABNORMAL shutdown with ' SYSTEM ERROR: latch 16 depth -1. (3715) ' and generating file named procore in users home folder.

2.My server Arch is Itanium 2 of HP rx3600 on 91D09.
May i have your valuable comments about this platform & progress 9.1D09?

3. Is progress version 9.1E04 certified for HP rx3600 i.e. Itanium Arch?

Thanks in advance
 
Thank you for your reply Mr.TOM.

1.We are not using 'KILL -9' .

IMHO 99.44% of all latch related errors are due to kill -9. You may not be using it but is it possible that someone else might be? I have seen:

1) Sys admins who don't talk to (or believe) their DBAs that are using kill -9. Sometimes behind the DBAs back.

2) Scripts, particularly orphan killers and idle user disconnect scripts that the admins either believed to be safe or, more than once simply forgot existed, be the root cause of these issues.

3) Help desks that either have a script or who manually use kill -9 to "fix" supposedly "hung" users.

4) Users who were just a little to knowledgeable about UNIX who kill their own sessions.

And lots of other use of kill -9 that is perhaps happening without your knowledge.

If your db is crashing with latch related errors it is almost always because someone, somewhere is using kill -9 (or an equivalent, kill -9 isn't the only signal with the "untrappable" property).

So it is worth putting time into double and triple checking every imaginable source to make sure.

So check. And then check again. In the meantime:

But still very frequently the DB goes for ABNORMAL shutdown with ' SYSTEM ERROR: latch 16 depth -1. (3715) ' and generating file named procore in users home folder.

Have you logged an issue and sent these to PSC tech support?

2.My server Arch is Itanium 2 of HP rx3600 on 91D09.
May i have your valuable comments about this platform & progress 9.1D09?

9.1d09 is ancient and unsupported. You should upgrade to 9.1e04.

Personally I'm not much of an HP fanboy although the "r" series servers were IMHO the best ones they've produced in the last 10 years.

3. Is progress version 9.1E04 certified for HP rx3600 i.e. Itanium Arch?

Not according to the Product Availability Guide

If it is not kill -9 (which I still place at a 99.44% probability) then it is a bug. Since you're running v9 your only recourse would then be to a) upgrade to 9.1e04 -- (which is 2 or 3 years old now) or b) upgrade to OpenEdge (10.1C is the current release).
 
Mr.TOM Thank you for your valuable response.

We have decided to upgrade Progress 10.1C from 9.1D09 for ia64 hp server rx3600 on HP-UX B.11.23.
Our application is QAD - M F G / P R O Release eB2 SP3 as of Jun 2 2003 for Manufacturing Management and it's works fine on another HP PA-RISC architecture systems.
If I upgrade to progress 10.1C, Will it works fine and is required any changes on my M F G / P R O Release eB2 SP3?

can you post few working sites on progress 10.1C with QAD- eB2 sp3?

Hence, i am extremely expecting your comments on this.

regards,

LENIN DASAN
 
Back
Top