In this DocumentGoal
Solution
Applies to:Oracle GoldenGate - Version: 9.5.0.5 and later [Release: 9.5.0 and later ]
HP NonStop S-series (Guardian)
Goal
=== ODM Question ===
How to resolve issue after a target HP NonStop system was coldloaded and source extract receiving error:
GGSEVT 103 EXTTMF Attempted to position past
current EOF on file \ODR2.$WEST4.TMFTRLS.AA000109, position
10872689, EOF 10866560, err -1
This can happen when a remote node is coldloaded or a serious system wide failure occuring. An extract writing via Expand to this failed remote node can receive this error after the system is back up and running.
Solution
=== ODM Answer ===
You will then need to do this for every extract that failed.
1) Stop the replicat(s) that is reading this trail.
2) Edit the replicat(s) param file to include HANDLECOLLISIONS
3) GGSCI > ALTER EXT <extract name>, begin 2010-11-30 19:00 (a timestamp that is some minutes before the failure)
Note: you can get a gauge of how much to go back by looking at the last record timestamp on the target node in the trail file using logdump.
4) GGSCI> ALTER EXT <extract name>, ETROLLOVER and START EXTRACT <extract name> 5) After a few minutes, when replicat is past the point of the time of failure, stop the replicat.
Remove HANDLECOLLISIONS and restart the replicat.