|
Just a little nitpicking: Western Digital do sell drives having more than 1 year warranty, look for the "RAID edition" drives. I think they have 3 or 5 years warranty, and also is "rated for 24/7 usage". I don't know if it's just snakeoil or if there is an actual quality difference. Also, for software RAID on linux, mdadm is the weapon of choice. Somewhat old Linux Software RAID HOWTO Linux Software RAID Wiki MDADM's main developer blog CeciPipePasPipe fucked around with this message at 17:18 on Mar 18, 2008 |
# ¿ Mar 18, 2008 17:14 |
|
|
# ¿ May 6, 2024 03:51 |
|
Also, here's another tip for Linux software raid: make sure you put your swap partition on RAID as well (instead of striping the swap partition). I learned this the hard way when one of the drives tanked and couldn't do a clean shutdown since it was unable to swap back in from the failed drive. Fortunately it didn't hurt anything important.
|
# ¿ Mar 18, 2008 17:40 |
|
admiraldennis posted:I agree, I would never put the operating system on a raid array meant for file storage. My current fileserver setup has 5x 500GB drives in a RAID-5 for storage and 2x 20GB fireballs in a RAID-1 for boot. If the storage array goes south, it's invaluable to still have a safely bootable system in order to diagnose things. Keeping boot and data separate is really a must in my opinion. I put the OS on a RAID1 partition across 4 drives (which is way overkill, but useful to keep identical partition tables - also then I could boot any of the four drives, in theory at least) and allocated the remaining parts of the disks to large RAID5 partitions (and some RAID1 swap again)
|
# ¿ Mar 18, 2008 20:54 |
|
What options are available if you want to add another 4 or 5 drives, and all the bays in the computer case are already taken? I guess adding a bunch of external USB drives is one way to go, but would probably result in subpar performance. Is it possible to buy some kind of external multidrive cabinet and hook it up to either a single eSATA port, or perhaps another SATA expansion card with external ports? I don't want anything that presents itself as one big pool of storage to the host, since I want to run Linux software RAID on the drives and also monitor individual drives' smart status.
|
# ¿ Apr 12, 2008 11:21 |
|
complex posted:http://www.newegg.com/Product/Product.aspx?Item=N82E16816855007 would probably work well for you. Ignore their RAID software and it will just be a JBOD box. Thanks for the tip
|
# ¿ Apr 13, 2008 15:26 |
|
Delta-Wye posted:and I won't fully understand it until I've spent money on parts and played with it some. It is a bit of a chicken and an egg situation, if you understand. Couldn't you just install a ZFS capable OS in a virtual machine and attach some small virtual hard drives to play around with right now? Or even just raid a couple of loopback-device-mounted files?
|
# ¿ Jul 1, 2008 21:52 |
|
porkface posted:Which virtualization software would you use for this? I've been trying to do a similar thing with physical disks in VMWare 1.x and it'll only recognize the first 128 GB of each drive. Why do you need such huge volumes or physical disks even? If you just want to play around with a RAID solution to get a feeling for how its tools work and how it handles failures, surely a couple of 1GB "drives" would be more than enough. In fact, before I sat up linux MD raid5 for real for the first time, I dd'ed in 5x10mb files from /dev/zero, attached them to the loopback file device thingies (man losetup) and assembled an array out of those.
|
# ¿ Jul 1, 2008 23:53 |
|
roflsaurus posted:Also, what do you guys use for automatic monitoring on headless boxes? I was thinking of writing a cron script to check drive status / usage, but also maybe other stuff like available apt updates, etc. If someone has one already written it would save me some time. mdadm and smartd configured to mail me and twitter me (-> free SMS warnings!) with errors and warnings; another cronscript that checks the contents of /proc/mdstat against a known hash as well for good measure. Also, subscribe to the debian-security-announce mailinglist for apt updates (or the ubuntu equivalent if you're on ubuntu).
|
# ¿ Jul 3, 2008 17:35 |
|
|
# ¿ May 6, 2024 03:51 |
|
Farmer Crack-rear end posted:For those people with huge arrays of SATA hard drives, in the event of a hard drive failure, how do you identify which hard drive is the one that needs to be replaced? Is there a way to set up the system that makes it relatively easy to find? The serial number is readily available via software, so on linux just make a note of the serial for each device, and when for example /dev/sdb goes down look up the serial number in your notes and compare with the label on the drive.
|
# ¿ Aug 4, 2008 19:56 |