This is Friday's work.

Gateway server of DGS network was rebooted to apply some package updates.
Downtime was from 13:11 to 13:13 JST on Oct. 10th.

-----
Although data transfer of LL h-of-t shouldn't depend on the DGS network, a lack of LL frames on Kashiwa cluster was coincident with that time ([1444104722, 1444104855)). I haven't concluded yet they are really correlated or not. Because power cycles of the gateway server of DGS is done only during the maintenance time in normal case, a lack of LL frames is not a serious matter from the view point of archiving data. On the other hand, a procedure of AR frame generation is slightly different depending on whether a chopped data itself or a chopped Science-mode. So this issue must be solved if they are really correlated.