Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
"Rich Braun" <richb at pioneer.ci.net> writes: > need a couple hundred gigs of storage. Note also that to run RAID5 > you need separate controllers for each drive. A standard > motherboard contains two PATA controllers, you need to make at least > a third available if you want good performance. (SATA is probably > the solution, now that the drives are nearly as cheap as PATA.) SATA is only barely more pricey than PATA. With 300G-400G drives (which is getting close to the minimum $/GB curve) the price differential between PATA and SATA is relatively small. You can get a PCI SATA driver with... 4 SATA ports? Or 8? Add that to the Mobo and you can probably get 6-8 SATA drives. With 8 400GB drives using RAID-5 + 1 hot spare you can combined them into 2.4TB of storage. My plan was to use this space for MythTV Backend Storage and as an online backup server for my desktops, laptops, and critical data backups (e.g. web server content). I'd like it all in a single file system instead of multiple partitions because I don't know a priori how much space I want to the different uses. >> But I also don't know how >> to combine the RAID vs. LVM vs. partitions in order to have >> a protected /boot partition in addition to a very very large / > > I don't bother with putting /boot on RAID, haven't found a reliable way to do > it yet. But it doesn't change very often so you can simply rsync it to your > other drive whenever you do a kernel update. I use RAID-1 on all my servers now. It's a cheap get-out-of-disk-failure-free card. I don't use LVM on those servers; I just don't see the point. It seems to add complexity to what I view as little gain. I have two MD devices, /boot and /. I've got grub installed on both drives and I've got grub.conf configured to boot off both drives. What this means is that I can pull out either drive and the machine still boots! I've got a script that will rebuild the raid array (including copying the partition table from the old drive to the new drive, setting up grub, and rebuilding the raid). It works great and takes all the guesswork out of a job that I rarely perform. Indeed, I've never (yet) had to perform this task on a live server; only in my testing. > The distros, at least SuSE, will set up LVM-over-RAID for you out of the box, > just follow the directions. Make sure to leave some unallocated space in the > LVM. (My advice--set up a small /boot partion, 200M to 500M, and make a > second partition /dev/md0 take up the rest of the space on the smaller of the > two drives. Make these two partitions the same size on the larger drive. If > the two drives are the same, proceed; if there is remaining space on the > larger one, you can set up some scratch storage. Then decide what logical I guess this was part of my question (and confusion).. Do I want LVM over RAID or RAID over LVM? Or do I want LVM over RAID over LVM? Also, if I want to do a RAID5 for / but RAID-1 for /boot, how do I want to lay that out? With RAID-5 do all my drives/partitions need to be the same size like they do with RAID-1? And then what's my upgrade path if I decide to swap out to larger drives? Let's say that 3-5 years from now I decide to swap out my 400G drives with 1TB drives -- what would be the right process to do that such that my main raid partition size increases? (Then I can resize my ext3 FS). I don't trust reiserfs for a server -- it's not designed to handle catastrophic failures like a power outage. Yeah, you could setup a UPS with auto-shutdown, but why take the extra chance with a fragile filesystem? -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord at MIT.EDU PGP key available -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |