On Feb 20, 2017, at 2:00 PM, Kirk Brooks wrote:
> I've trimmed out the declarations and such for clarity and because I don't
> think they are the issue. This is a method that gets called a lot. I dare
> say several hundred times a day, probably more. But I keep seeing 1 or 2
> errors. Here's a sample error report:
>
> Partner_get_name :
> [db engine] Invalid table number requested by a Plug-In
> Error # -1
> Line # 51
> Process :11 [ORDER_EDIT_RECORD]_ORDERS_124045
> Environment:[RM C]
> Session:
> {"id":32044,"userName":"***","4dUser":"***","dbUser":-26,"
> userId":4048,"machine":"***","machineName":"***_sched","partnerId":130646,"
> boundPartnerId":101,"boundVendorId":101,"salesRepId":4177,"registerId":
> "***_32044"}
> Error Stack:
> 1002 : SQCO; Failed to connect to the remote server.
> 1009 : SQLS; Remote SQL Server is not available.
>
>
> There's nothing in the info from 4D that helps me. The table is clearly
> valid. But I'm not entirely clear what the SQL errors are trying to tell
> me. I've attempted to recreate the error by feeding the method various
> levels of junk but I can't get it to barf.
>
> Any ideas?
The Error Stack caught my eye. To me it indicates 4D Server, for whatever
reason, did not respond in an expected way. Maybe it was not quick enough.
Maybe 4D Server and the SQL Server was a bit busy at that exact point in time
and thus returned error 1009 which then cascaded to error 1002.
You said this code is called many hundreds of times each day, yet occasionally
if fails. So the program code is good and passing it bad data in $partner_id
should not cause an error like this.
I would suggest you write a wrapper method for this code and check for the 1002
or 1009 error code returned. If you get that, just wait a bit and try again.
Maybe try 3 times before you give up.
Maybe do a PAUSE PROCESS(Current process;10) to wait 1/10 of a second, then run
the code again and if there is still an error, repeat. Keep an error count so
when you reach your limit of 3 attempts, then you record the error to your
error logging and to the user.
I would also record a special info type message every time you run into this
situation where you have to try more than once and record success or failure
and the number of attempts. Then you’ll be able to track how often this
fix/hack is working and being successful. You might find you need to try 5
times before it is always successful.
I had to implement a similar fix/hack for a situation where I was checking if a
shared folder was mounted using the “Test path name” command. Sometimes it
would report the share was not available, but it really was. If you did the
same operation a second time it worked. By simply checking several times in a
row if the first Test path name failed solved the problem and eliminated the
user seeing any error message.
For anyone interested, this error can occur when you are using macOS 10.7 to
share a directory with a Windows user via SMB. Apple threw out SAMBA as their
Windows SMB code base and wrote their own from scratch with 10.7. Whatever they
did caused this issue for me. Might have been fixed in later macOS versions.
Tim
********************************************
Tim Nevels
Innovative Solutions
785-749-3444
timnev...@mac.com
********************************************
**********************************************************************
4D Internet Users Group (4D iNUG)
FAQ: http://lists.4d.com/faqnug.html
Archive: http://lists.4d.com/archives.html
Options: http://lists.4d.com/mailman/options/4d_tech
Unsub: mailto:4d_tech-unsubscr...@lists.4d.com
**********************************************************************
