When mirroring using pget, if a big file being downloaded got interrupted because of network/connection problem,
the mirror command thinks the big file has already complete download when actually is not (XXX.pget.status still present),
Thus it changes the timestamp of the incomplete file to that of returned by server. So next time when you issue the mirror --continue
to resume the mirror, lftp incorrectly thought the big file as complete download and delete its corresponding pget.status, leaving
a broken, incomplete file that is hard to spot/find.
When mirroring using pget, if a big file being downloaded got interrupted because of network/connection problem, the mirror command thinks the big file has already complete download when actually is not (XXX.pget.status still present), Thus it changes the timestamp of the incomplete file to that of returned by server. So next time when you issue the mirror --continue to resume the mirror, lftp incorrectly thought the big file as complete download and delete its corresponding pget.status, leaving a broken, incomplete file that is hard to spot/find.