Classify disk/tape errors for TAS
We should improve TAS/end of tape session message to lessen the impact of errors that are not caused by the tape infrastructure. See https://gitlab.cern.ch/cta/operations/-/issues/928 for details.
We already have something similar: in DiskWriteTask
, TapeReadSingleThread
, TapeReadTask
, etc, the watchdog remembers the type of the error. For example:
currentErrorToCount = "Error_diskOpenForWrite";
currentErrorToCount = "Error_tapeUnload";
if (!currentErrorToCount.empty()) {
m_this.m_watchdog.addToErrorCount(currentErrorToCount);
}
However, this is not added to the log messages.