Open nikmagini opened 8 years ago
Currently, the replicas of files in blocks reactivated by the BlockActivate agent are always created in status 'not staged' (t_xfer_replica.state=0).
Any subsequent replica on Disk nodes is created with status 'staged' (t_xfer_replica.state=1) by FilePump instead
This means that the FileRouter will be biased in selecting subsequent replicas, even if the original replicas would be faster.
Need to apply a fix similar to what was done for the Inject API in https://github.com/dmwm/PHEDEX/issues/1024
Currently, the replicas of files in blocks reactivated by the BlockActivate agent are always created in status 'not staged' (t_xfer_replica.state=0).
Any subsequent replica on Disk nodes is created with status 'staged' (t_xfer_replica.state=1) by FilePump instead
This means that the FileRouter will be biased in selecting subsequent replicas, even if the original replicas would be faster.
Need to apply a fix similar to what was done for the Inject API in https://github.com/dmwm/PHEDEX/issues/1024