Allbase connect DBERR 14031

[Ian Little]

When we try to connect to an Allbase database we get a DBERR 14031. Any ideas on how to overcome this problem?

[Mei Jiao]

Hi,

You may find this document helps:
http://itrc.hp.com/cki/bin/doc.pl/screen=ckiSearchResults?mode=id&searchString=KBRC00009784

Do keep me posted.

[Werner Ittner]

having too many allbase processes running can cause shared memory problems. (Especially because daemons still run about 30 secs after job has been finished).

I had this problem as well. HP didn't find a solution, then.

What I do is to controll shared memory access myself in a shell programm.

I check number of SM accesses.

(ipcs -m ....)

only if number of these jobs is smaller than a predifined amount (depending on your hardware and kernel parameters) then the next allbase programm is being fired.

We run lots of allbase systems with lots of users and since then no problems anymore.

Werner

[Ian Little]

To explain more the problem arose when we had to reboot the machine due to a scsi problem. We were unable to take our processes down cleanly and hence we ended up with the allbase problem.

The problem was solved with the following answer from HP:

***DO NOT TRY THIS UNLESS IT IS YOUR LAST RESORT ****

this procedure reset's the address of where the shared memory segment should start.

Obviously it is not a first resort. the ipcs process should be followed before trying this.


Where /hertznet/allbase/htznet2/HTZNET1 is the path to your dbcon file.


# adb -w /hertznet/allbase/htznet2/HTZNET1

1a4?X <----- type this, no prompt in adb

1A4: 808 <----- adb returns this location and contents

1a4?W -1 <----- type this, offset, write and replace with -1

1A4: 808 = -1 <----- adb returns this >!]