Hello all,
we are getting some ATLAS tests that transfer data from our DPM nodes
but they sometimes fail during the transfer.
The reason (according to dpm and gridftp logs) is that the transfer
takes too long and so the gridftpd receives abort request.
Sometimes the abort comes after one hour, sometimes after 4 hours.
Is this an attribute of the appropriate FTS channel?
Is this abort issued only after fixed time or is there more
sophisticated algorithm (e.g. based on transfer speed over last 10 minutes)?
We are not aware of any problems with machine or network load, so
it is a bit difficult to say what is the cause of slow transfer.
Thank you for any hint,
--
Tomas Kouba
Institute of Physics, Academy of sciences of the Czech Republic
|