Our current Pool and JDC update the nTime field of every job using the current system time, instead of using the value sent by TP. This leads to mine invalid blocks when the value is too old with respect to the network one. We were able to notice this bug on testnet4, since its MTP is moved in the future.
Here some logs which shows TP logs when this bug is in place:

cc @Sjors
Our current Pool and JDC update the
nTimefield of every job using the current system time, instead of using the value sent by TP. This leads to mine invalid blocks when the value is too old with respect to the network one. We were able to notice this bug ontestnet4, since its MTP is moved in the future.Here some logs which shows TP logs when this bug is in place:
cc @Sjors