I am hoping for some help to understand what edge condition i seem to be encountering with my Drobo Pros. They are connected via a shared ethernet segment and accessed via iSCSI from CentOS / RHEL 5.5.
I can copy on and off flat out with out any issues, but when i run a read / write load using pbzip2 (parallel bzip2) i get detected conn errors, and unchecked the drobos power down (or at least stop responding over iSCSI and the lights go out). Drobos are two LUNS each formatted to ext3 / 8TB each.
Anyone with any thoughts at to how i might measure what the difference is between pbzip2 or a straight cp command, and / or if i can get crash logs of the drobo devices directly?
Many Thanks For Reading,
P.S. This effects all four of my 1.1.5 drobo pros connected to either my desktop or server machines (both running CentOS 5.5)