FUJ00130827
FUJ00130827
HORIZON KEL dsed5628Q (Version 4)
Type: Unresolved
Status: Deprecated
Title: Stockunit loss / gain not sent to local suspense
Summary: Stockunit loss / gain not sent to local suspense when stockunit was rolled during CABSProcess
Raised: by Dave Seddon on 10/01/2008
Last updated: by Anne Chambers on 01/08/2008
Release: s70
System product: EPOSS
Keywords: local suspense suspence bts non zero
Status: Deprecated
Visi Medium
Peak: PC0152376
TFS: 782747
Symptoms
When a stockunit was rolled over at 7pm the loss/gain for that stockunit was not sent to local suspense and consequently when local
suspense was cleared this loss/gain was not included. The value of this loss/gain was shown on the trading position line on the branch trading
statement which would always be zero if there was no problem. <br><br>In the counter application event log the following pair of events were
written at 7pm in between events that showed the CABSProcess was running...<br><br>Source: Riposte<br>Text: An unexpected error
occurred while attempting to modify an entry in the run map. Timeout occurred waiting for lock.<br><br>Source: EPOSSStockUnit<br>Text:
Warning Messaeg: Unexpected error executing fPostTxnsToLocalSuspense - see audit log for details. <br> <br>In the counter audit log the
following was written at the time...<br><br>SU:fPostTxnsToLocalSuspense (:- 1056374781) Timeout occurred waiting for lock. (0xC1090003)
CreateMessageEx: RiposteCreateMessageEx call failed.
Problem
The probiem is that because of a previous Peak, PC0140715, CABSProcess writes out messages atomically. It does a StartTransaction quite
early on (which creates the lock), then initiates writing lots of transactions with CreateMessage and persistent objects with PutObject and
finally really writes them with a call to EndTransaction (which ends the lock). If something else tries to write a transaction whilst CABSProcess
has things locked then it will time out after 10 seconds. Hence if CABSProcess takes more than 10 seconds to run you could get this sort of
problem. In the first case seen, PC0152376, the CABSProcess took 33 seconds to run which gave a significant window of opportunity for this
sort of problem to occur.<br><br>Note: The CABSProcess only runs on the gateway counter so the problems seen will not occur on slave
counters.
Solution - Helpdesk
No fix planned for Horizon given the relative rarity of the problem. However, should the problem start occurring more often then the need for a
fix should be reviewed. Add any cases to list below. <br><br>MSU will need to be informed of what has gone wrong and what we believe needs
to be done to correct matters. We need to look at impact on counter as well as the impact on POLFS. See PC0152421 as an example of what
was done in one particular case but be aware that corrective actions in particular will not always be the same and these will need to be
looked at on a case by case basis.<br> <br><br> List of cases seen...<br><br>Date Peak Reference<br> 12/12/2007 PC0152376 - problem with
local suspense<br>05/03/2008 PC0155120 - DEF did not rollover auto with the Office 27/12/2007 PC158102 - problem with local suspense
Evidence
Event log Audit log Messagestore