Re: Software RAID and drive failures

2003-03-12 Thread Kirk Strauser
At 2003-03-12T07:03:59Z, Ron Johnson <[EMAIL PROTECTED]> writes: > 1. Tagged command queuing should order the reads. Yes, but you still have contention in that a larger-than-normal chunk will be read from file1, alternating with a larger-than-normal chunk from file2. > 2. As a DBA, it seems to m

Re: Software RAID and drive failures

2003-03-11 Thread Ron Johnson
On Tue, 2003-03-11 at 14:30, Kirk Strauser wrote: > At 2003-03-11T18:41:12Z, Ron Johnson <[EMAIL PROTECTED]> writes: > > > Why wouldn't you get parallel reads, especially if you do async IO? > > I guess a better example would be when two files are accessed, and each > has blocks that physically r

Re: Software RAID and drive failures

2003-03-11 Thread Kirk Strauser
At 2003-03-11T18:41:12Z, Ron Johnson <[EMAIL PROTECTED]> writes: > Why wouldn't you get parallel reads, especially if you do async IO? I guess a better example would be when two files are accessed, and each has blocks that physically reside on the same two drives. Imagine that both files are acc

Re: Software RAID and drive failures

2003-03-11 Thread Ron Johnson
On Mon, 2003-03-10 at 13:15, Kirk Strauser wrote: > At 2003-03-10T18:25:06Z, Ron Johnson <[EMAIL PROTECTED]> writes: > > > I bet RAID0 would be even faster. Yes, yes, fragile... > > Not for concurrent reads. If a file spans more than one physical disk, then > at least two drives have to be read

Re: Software RAID and drive failures

2003-03-10 Thread Alvin Oga
On 10 Mar 2003, Ron Johnson wrote: > On Mon, 2003-03-10 at 09:22, Kirk Strauser wrote: > > At 2003-03-10T13:10:02Z, Ron Johnson <[EMAIL PROTECTED]> writes: > > > > > ?? RAID1 can only handle 1 failure also!! > > > > Depends on how many drives are in the volume, doesn't it? > > Hmmm. By d

Re: Software RAID and drive failures

2003-03-10 Thread Kirk Strauser
At 2003-03-10T18:25:06Z, Ron Johnson <[EMAIL PROTECTED]> writes: > I bet RAID0 would be even faster. Yes, yes, fragile... Not for concurrent reads. If a file spans more than one physical disk, then at least two drives have to be reading from the same region at the same time. With RAID1, you ca

Re: Software RAID and drive failures

2003-03-10 Thread nate
Ron Johnson said: > Hmmm. By definition, RAID1 is "mirroring". Never heard of putting any > more than 2 disks in a mirrorset. Never even thought of it, > actually. With more than 2 disks, I'd automatically go with RAID5. in many cases I would too. However, most implimentations of LILO and or

Re: Software RAID and drive failures

2003-03-10 Thread Ron Johnson
On Mon, 2003-03-10 at 11:36, Kirk Strauser wrote: > At 2003-03-10T16:39:36Z, Ron Johnson <[EMAIL PROTECTED]> writes: [snip] > > Never heard of putting any more than 2 disks in a mirrorset. > > Mirroring can have a nice boost for read access. I'd imagine that such a > setup would be useful for som

Re: Software RAID and drive failures

2003-03-10 Thread Kirk Strauser
At 2003-03-10T16:39:36Z, Ron Johnson <[EMAIL PROTECTED]> writes: > Hmmm. By definition, RAID1 is "mirroring". Correct. > Never heard of putting any more than 2 disks in a mirrorset. Mirroring can have a nice boost for read access. I'd imagine that such a setup would be useful for something li

Re: Software RAID and drive failures

2003-03-10 Thread Frank Gevaerts
On Mon, Mar 10, 2003 at 10:39:36AM -0600, Ron Johnson wrote: > On Mon, 2003-03-10 at 09:22, Kirk Strauser wrote: > > At 2003-03-10T13:10:02Z, Ron Johnson <[EMAIL PROTECTED]> writes: > > > > > ?? RAID1 can only handle 1 failure also!! > > > > Depends on how many drives are in the volume, does

Re: Software RAID and drive failures

2003-03-10 Thread Ron Johnson
On Mon, 2003-03-10 at 08:24, Massimiliano Ferrero wrote: > > ?? RAID1 can only handle 1 failure also!! > > > > If you are a decent administrator, you'll act when a disk becomes > > flaky, or as soon as it fails. (Of course, if 2 disks go at the > > same time, you're hosed, but that's the cas

Re: Software RAID and drive failures

2003-03-10 Thread Ron Johnson
On Mon, 2003-03-10 at 09:22, Kirk Strauser wrote: > At 2003-03-10T13:10:02Z, Ron Johnson <[EMAIL PROTECTED]> writes: > > > ?? RAID1 can only handle 1 failure also!! > > Depends on how many drives are in the volume, doesn't it? Hmmm. By definition, RAID1 is "mirroring". Never heard of putt

Re: Software RAID and drive failures

2003-03-10 Thread Kirk Strauser
At 2003-03-10T13:10:02Z, Ron Johnson <[EMAIL PROTECTED]> writes: > ?? RAID1 can only handle 1 failure also!! Depends on how many drives are in the volume, doesn't it? -- Kirk Strauser In Googlis non est, ergo non est. pgp0.pgp Description: PGP signature

Re: Software RAID and drive failures

2003-03-10 Thread Mike Dresser
On 10 Mar 2003, Ron Johnson wrote: > ?? RAID1 can only handle 1 failure also!! Well, actually you could have a N>2 number of drives in your raid-1 setup. It's rarely done, but still possible. Mike -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Co

Re: Software RAID and drive failures

2003-03-10 Thread Massimiliano Ferrero
?? RAID1 can only handle 1 failure also!! If you are a decent administrator, you'll act when a disk becomes flaky, or as soon as it fails. (Of course, if 2 disks go at the same time, you're hosed, but that's the case for RAID1 also...) Hmmm, probably I wasn't very clear. What I mean here is

Re: Software RAID and drive failures

2003-03-10 Thread Ron Johnson
On Sun, 2003-03-09 at 16:48, Massimiliano Ferrero wrote: > Calyth wrote: > > I'm no expert comparing to the other people here who've experimented the > > whole thing out, but I'd think that you might want to get at least 3 IDE > > drives and set up RAID 5 instead. > > I'm no expert either nor I

Re: Software RAID and drive failures - booting

2003-03-10 Thread Calyth
I didn't know about the part that RAID 5 isn't good for small writes, but after a little thought it does make sense. I'm fairly aware to make Linux in general to boot from a software RAID 5 is quite a pain, maybe I should've mentioned to put the /boot in a small drive that isn't RAID 5, I recall

Re: Software RAID and drive failures

2003-03-09 Thread Massimiliano Ferrero
Calyth wrote: I'm no expert comparing to the other people here who've experimented the whole thing out, but I'd think that you might want to get at least 3 IDE drives and set up RAID 5 instead. I'm no expert either nor I want to start a religious war ;) but... RAID 5 had much more sense when the

Re: Software RAID and drive failures - booting

2003-03-09 Thread Alvin Oga
> Calyth said: > > I'm no expert comparing to the other people here who've experimented the > > whole thing out, but I'd think that you might want to get at least 3 IDE > > drives and set up RAID 5 instead. > > Say you got 3 20GB HDs, RAID 1 (mirroring) would only give you 20GB of > > space, whil

Re: Software RAID and drive failures

2003-03-09 Thread nate
Calyth said: > I'm no expert comparing to the other people here who've experimented the > whole thing out, but I'd think that you might want to get at least 3 IDE > drives and set up RAID 5 instead. > Say you got 3 20GB HDs, RAID 1 (mirroring) would only give you 20GB of > space, while RAID 5 would

Re: Software RAID and drive failures

2003-03-09 Thread Calyth
I'm no expert comparing to the other people here who've experimented the whole thing out, but I'd think that you might want to get at least 3 IDE drives and set up RAID 5 instead. Say you got 3 20GB HDs, RAID 1 (mirroring) would only give you 20GB of space, while RAID 5 would provide more, yet s

Re: Software RAID and drive failures

2003-03-09 Thread Massimiliano Ferrero
nate wrote: Juhan Kundla said: I buy two smaller (and cheaper) IDE disks and use them in RAID-1 array. I hope that this gives me good protection against hardware failures. If one disk fails, then other will still have my data intact, right? The main question is, that how good is the software RAID

Re: SmartUPS (Re: Software RAID and drive failures)

2003-03-08 Thread nate
Joao Pedro Clemente said: > Hi. Sorry for bugging in, but I got curious: > These SmartUPS really do "wave-correction"? I've heard, from other brand > vendors (some Italian brand) that APC didn't had ANY such UPS, that they > only worked in a garanteed sine-wave with correct voltage when operating

SmartUPS (Re: Software RAID and drive failures)

2003-03-08 Thread Joao Pedro Clemente
> the system operates in an ideal enviornment with 90CFM of airflow > going through the case, a sine-wave APC SmartUPS powering it, as Hi. Sorry for bugging in, but I got curious: These SmartUPS really do "wave-correction"? I've heard, from other brand vendors (some Italian brand) that APC didn't

Re: Software RAID and drive failures

2003-03-08 Thread Alvin Oga
hi ya On Sat, 8 Mar 2003, Juhan Kundla wrote: > > I have a task to set up a file server. I have very small budget but > quite high demands for data integrity. I must do everything possible to > ensure, that i can recover data after hardware or user failures. > The performance of the computer is

Re: Software RAID and drive failures

2003-03-08 Thread nate
Juhan Kundla said: > I buy two smaller (and cheaper) IDE disks and use them in RAID-1 array. I > hope that this gives me good protection against hardware failures. If one > disk fails, then other will still have my data intact, right? The main > question is, that how good is the software RAID, whe

Re: Software RAID and drive failures

2003-03-08 Thread Ron Johnson
On Sat, 2003-03-08 at 13:40, Juhan Kundla wrote: > Hei! > > I apologise, that i bother you with questions, that are not entirely > Debian specific. I don't know any better place to ask those questions. > If someone here could point a newbie like me to right direction, it > would be very much appre

Software RAID and drive failures

2003-03-08 Thread Juhan Kundla
Hei! I apologise, that i bother you with questions, that are not entirely Debian specific. I don't know any better place to ask those questions. If someone here could point a newbie like me to right direction, it would be very much appreciated. I have a task to set up a file server. I have very s