What would happen if the CPU bottlenecked? Would I be looking at the NAS's OS freezing up completely or would data transfer rates just cap out at a certain speed? Working on putting some of it together now. You make an interesting and good point about CPU being a possible bottleneck.
When I was asking about the speed of PCIe GEN1 I was more referring to would it support the max 6gbps SATA speeds of the LSI 9260 card? I have to upgrade the storage controller from P400 to something that supports 6gbps but I want to make sure that it will actually be able to handle those speeds. If I could achieve 2-6gbps performance it would be fine. I plan on using it to store 4K video footage so I want fast speeds hence the 10GbE but it doesn't have to be THAT fast. Was considering putting in a small PCIe->NVMe adapter so I could have a cheap NVMe cache but honestly I don't know if I will need it for the 1-2 people who would be accessing it at any given time. Im considering stuffing the server full of 2TB cheap laptop SSD's drives and either running FreeNAS or possibly vSAN. Maybe clear out the cobwebs.Ĭould be sitting on a big issue (or not) but i'd take precaution.Hey thanks for the helpful info.
I'd get a damn good backup of that machine, SPP patch up everything to the latest firmware and take a good look for any issues in the integrated logging or ilo event log. It was the least logical but most obvious given how tight hp fits their servers.
Also some asshole that built mine used too long of a battery wire, which when the battery reached a certain aged, caused a power issue, killing the cache boards endlessly until the cable was replaced with the right length. Proliant D元80 G5 Options - SATA 2.5' (SFF) Multi-Level Cell (MLC) Solid State Drives (SSDs) Proliant SATA Hot-Plug 2.5' (SFF) Multi-Level Cell (MLC) Solid State Drives (SSDs) - Option Part Numbers 636597-B21 HP 400GB 3G (3Gb/s) Hot-Plug Serial ATA (SATA) 2. If the cache board dies, it takes out a drive and marks itself bad, the p400 will then not operate if you take the bad cache board off. Speaking of which if the cache board or battery faults the p400 does all kinds of sketchy shit. After 3 years the battery for your p400 write cache is hosed anyways.
in the array pauses for 1 or 2 seconds while the new drive is initializing.
You know you can get 30% more iops or more by replacing that p400 array say with a p410/fbwc. Title: HP Smart Array Controllers/HP ProLiant Servers - How to Replace a. I've spent a lot of time with these servers eh :) (ole pull and plug back in trick without user intervention).Īlso patrol scan for bad sectors defaults to 3 seconds, but can go to 0(higher than read) and i've chosen 1 before to seek out bad sectors and remap them whilst not busy - which can lead you to think it's rebuilding when it's not busy. A loose connection perhaps could cause 50% of the drives to drop, and HP is wise enough to then auto-rebuild. So in theory you could lose 50% of the drives and continue operating. Proliant SATA Hot-Plug 2.5' (SFF) Midline (MDL) Drives- Option Part Numbers. Proliant D元80 G5 Options - SATA 2.5' (SFF) Midline (MDL) Hard Drives. It is very well possible to lose an entire port(cable or drive chassis), power to such, or failure of the raid card and continue to operate assuming sufficient remaining drives (ie use raid-1+0, even raid-1 is raid 1+0 on hp for expansion sake). 460426-001 HP 250GB 3G (3Gb/s) Hot-Plug Serial ATA (SATA) 5.4K 2.5' (SFF) Entry (ETY) Hard Drive. Remember D元80's are dual ported (with dual ported drives) and have redundant cages, (could be g6/g7) - so in a raid-1+0 pair the system does not put the adjacent drives next to each other.