Recent content by bonox

there are dud ones out there - possibly the drives aren't seated fully, the interposers are dead (you are using interposers right? You haven't shown a picture of a drive in the sled with the interposer fitted) or there's a board failure in the circuit somewhere. Just of note, straining to look...

i think i'm being trolled :rolleyes:

they don't - take another look at the cable picture above - the sas connector on the cable end goes to the drive. THe blue lines go back to the hba. The sata power connector goes to the back of the drive connector where you see alll the copper fingers

the equivalent of your lsi card connector on one end plus this will let you use that sata power connector

Background: There's an A controller and a B controller. I think they're labelled as such by arrows on the left side of the picture you posted in the middle of the box. Each controller is connected to one of the two signal paths to a SAS HDD. SATA has only one path, so will be connected only to...

I didn't recommend any card, but i'm very happy with my older pcie2.0 (ie 9200 type cards). I personally run 60 3.5" disks and 25 2.5" disks in 5 shelves off one of those old cards and since I don't run mirrors everywhere, i'm much more constrained in IOPS than card/bus bandwidth. You should...

the sff 8088 cables you'll be using carry four sas lanes, meaning 24Gb/s. If you connect that (singly or via daisy chaining) to one or more shelves, then that 24Gb/s is shared by however many drives are hanging off it. (45 in your original proposal) If you have four ports, you've got 16 sas...

that card has 8 channels across the two ports, so 48Gb/s. If you only use one port you're still at 24Gb/s aggregate capacity. You can choke it artificially if you create pools that manage to use only one channel, but if you're using everything, you'll see much more than 6Gb/s. You're likely to...

unless you've got monetary issues with downtime I wouldn't go for multipathing, especially with a copy on write filesystem. And if I did have hefty SLA's, I go for clustering first since you cover the server going down in addition to the data path. get rid of the HBA to top shelf link as well...

unless you've got software that support multi-pathing, the right side of the connection tree is pointless. The Cards will say in their literature something like "Supports up to 254 devices" or some such, which means keep adding shelves up to that point, after which that's what you'd use the...

lots of these drives are used in non zfs applications (netapp or SAN's in general) where they are dumb, and most were probably intended to sit behind raid controllers rather than software control. May also be much more relevant to block level storage rather than file level, but that's my...

for an individual drive, it's irrelevant. I believe it relates to patrol reads/scrubs on arrays where the file system will complain of an error if it compares blocks from different drives. If the blocks are free, they should all have the same 'data' right! Once a block is marked free for use (ie...

indeed i was :)

art of server (youtube) has a video or two about how to do it, notably on the 2x2 port cards built with a pcie switch cards. The process is identical for the single cpu cards like yours. Additionally, he notes common issues with a port not working - some seem to be related to simple card...

i used an ordinary pc, but had to use the 5/6 pin tape trick to avoid boot looping