meanwhile i want to provide an update on this, i i did a citation "it would be totally viable". that's not true.
>i have HP DL380e Gen8 with P420 and did thoroughly test it in passtrough mode with ZFS and 12 SATA disks , i.e.
>lots of IOPs, bandwidth saturation with dd read/write of uncomressed data streams, scrub etc - i cannot see any
>problem, it works rock solid for me.
>Just wanted to throw it out there that the p420i at least is a totally viable HBA.
it seems to really depend which disk you use and what type of errors occur.
based on recent experiences with my p420i in hba mode while thoroughly testing/wiping older harddisks and sorting out old/bad disks for trash , i can tell this hba (compared to lsi) really sucks regarding error handling. while doing nothing but "dd" to or from several disks concurrently , i had multiple hangs of the whole controller and i/o stopped completely to all disks.
that would explain why i had at least two or three complete lockups with zfs when using that controller in a test server a while ago.
also see
https://community.hpe.com/t5/ProLia...-Gen8-P420i-Controller-Lock-Up-s/td-p/7059791
conclusion:
p420i in hba mode - use at your own risk and don't use it in production. test thoroughly and test with bad disks.