[Lustre-discuss] [URGENT] lustre 1.8.1 - data loss while failover

[簡嘉宏]

Hi,

   I am trying the lustre failover mechanism while wrf program running and dumping data by each time snapshot on lustre filesystem. It seems that while failover, it will loss one dump and getting the error message from wrf like this:
    open_hist_w : error opening wrfout_d01_2009-05-18_11:00:00 for writing ***
   I think it was the code problem because of time out reached, then the program will give up this dump. This is my understanding.

   Any idea or other suggestion ?

Thanks a lot!


Nelson