elkay14

I have 5 drive sleds from a Norco SS-500 that died (backplane failure.) They are the newer style and don't fit my older SS-500. Free to anyone who wants them, just cover the shipping.

It's hard to help you without more information. How many drives are attached to the 640L? They just don't appear in the device list? Perhaps give us output of lspci and dmesg

You probably want this one if you get a v1 or v2 chip: http://www.asrockrack.com/general/productdetail.asp?Model=EPC602D8A#Specifications 2 x PCIE3.0 x16, 4 x PCIE3.0 x8, 1 x PCIE2.0 x8

Ok, it is back. I had to edit disk.cfg to not start the array automatically and then edit the cache fs type back to xfs. This is definitely something that should be addressed in the UI. I shouldn't be able to change the config such that the UI is locked up.

I can't change the slots back to 1... emhttpd is hung up

My cache drive is xfs, so why is it complaining about btrfs? This cache drive was once btrfs, but that was months and months ago. I think this may be a bug?

Ok, done...

I stopped my array, changed the number of array devices to 12 and then changed cache devices to 4. The admin page is stuck at "Mounting disks..." I see in the syslog: Jan 11 07:58:45 arwen emhttp: shcmd (169): mkdir -p /mnt/cache Jan 11 07:58:45 arwen emhttp: mount error: No file system (no btrfs UUID) Jan 11 07:58:45 arwen emhttp: shcmd (170): rmdir /mnt/cache Jan 11 07:58:45 arwen emhttp: shcmd (171): sync Jan 11 07:58:45 arwen emhttp: shcmd (172): mkdir /mnt/user0 Jan 11 07:58:45 arwen emhttp: shcmd (173): /usr/local/sbin/shfs /mnt/user0 -disks 1022 -o noatime,big_writes,allow_other |& logger Jan 11 07:58:45 arwen emhttp: shcmd (174): mkdir /mnt/user Jan 11 07:58:45 arwen emhttp: shcmd (175): /usr/local/sbin/shfs /mnt/user -disks 1023 2048000000 -o noatime,big_writes,allow_other -o remember=0 |& logger Jan 11 07:58:45 arwen emhttp: shcmd (176): cat - > /boot/config/plugins/dynamix/mover.cron <<< "# Generated mover schedule:#01230 0 * * * /usr/local/sbin/mover |& logger#012" Jan 11 07:58:45 arwen emhttp: shcmd (177): /usr/local/sbin/update_cron &> /dev/null So what now?

If you have an HP microserver with the original 2GB in it, here's your chance. I upgraded my N40L to 8GB a couple years ago and I've got the original 2GB DIMM to give away (early spring cleaning in progress.) Free to anyone who has an HP microserver. It goes to the first person who posts a pic of their microserver including a note that says '2GB please!'

These are $200 at Newegg today...

It appears there is a bug that keeps this card and any other Marvell 88SE923X based card from working properly in Linux. It is likely that there is some AHCI implementation weirdness on Marvell's part that is triggering bad behavior in the AHCI driver, causing the controller to freak out under load and a SMART command being issued. Stay away, as it doesn't look likely that this will be fixed in the kernel anytime soon. My RocketRAID 640 card does not use one of these chips -- there are two 9128 chips onboard instead.

I was mistaken; I have a RocketRAID 640... not the L I'll do some more research on this L model.

What hardware version do you guys have? I have had one of these in my 4.7 install for over a year without issues.

Sorry - my VM running the site was down. Try again!

Since there is a kernel driver loaded for it, you won't need to do anything else.