[linux] raid question

Mark L. Chang mlchang at u.washington.edu
Wed May 7 22:35:10 PDT 2003


Well, it could really slow things down. That's what I saw in my case.

On Wed, May 07, 2003 at 10:28:27PM -0700 or thereabouts, A. Mathes wrote:

> it's cool to use the drives while it's syncing, right? It's not like

> defrag or something where it has to start over every time?

>

> Andrew Mathes

>

> On Wed, 7 May 2003, Mark L. Chang wrote:

>

> > Shouldn't take that long. Mine got 'stuck', but that was a fresh install

> > of rh7.3 that I just redid. Had no idea what happened, just redid it and

> > it went okay.

> >

> > Perhaps it is something to do with the IDE controllers you have. Are

> > they on the same channel? They all should be on separate channels

> > (master on all of them). It seems like two of yours are on the same

> > channel, maybe I'm not reading that right. That can slow things down

> > tremendously.

> >

> > Maybe there is something with the partitions being wrong? Check the logs

> > -- any IDE errors?

> >

> > On Wed, May 07, 2003 at 05:31:48PM -0700 or thereabouts, A. Mathes wrote:

> > > know why it's taken 24 hours on a 2xp3 933 1 Gb ram ... 3 120 Gb HD's? I

> > > read somewhere about a min/max setting for reconstruction so it

> > > wouldn't eat up other processes or something...

> > > it's only going at 1229K/Sec ... and nothing else is running.

> > >

> > > Andrew Mathes

> > >

> > > On Wed, 7 May 2003, Mark L. Chang wrote:

> > >

> > > > When you set it up initially, there isn't anything but crap on the three

> > > > drives. It has to put them into a known consistent state, and that's the

> > > > rebuild process. Looks just like when a drive horked.

> > > >

> > > > It takes some time. It took just under 2 hours on my Athlon 1GHz machine

> > > > with 3x80gb 8mb cache drives at ATA-133.

> > > >

> > > > On Wed, May 07, 2003 at 05:11:45PM -0700 or thereabouts, A. Mathes wrote:

> > > > > so, i set up a server this weekend using 3 hd's. It wouldn't boot off

> > > > > RAID-5 for /boot partition, so, I ended up setting up /boot as RAID-1 and

> > > > > / as RAID-5.

> > > > > It seems like everything is going well, except when I cat /proc/mdstat:

> > > > >

> > > > > Personalities : [raid1] [raid5]

> > > > > read_ahead 1024 sectors

> > > > > md1 : active raid1 ide/host0/bus1/target0/lun0/part1[2]

> > > > > ide/host0/bus0/target1/lun0/part1[1] ide/host0/bus0/target0/lun0/part1[0]

> > > > > 96256 blocks [2/2] [UU]

> > > > > resync=DELAYED

> > > > > md2 : active raid5 ide/host0/bus1/target0/lun0/part3[2]

> > > > > ide/host0/bus0/target1/lun0/part3[1] ide/host0/bus0/target0/lun0/part3[0]

> > > > > 236010752 blocks level 5, 32k chunk, algorithm 3 [3/3] [UUU]

> > > > > [=========>...........] resync = 47.6% (56217208/118005376)

> > > > > finish=837.0min speed=1229K/sec

> > > > > unused devices: <none>

> > > > >

> > > > > well, it's resyncing, to say the least. What's confusing me is I had

> > > > > thought resync happened when a drive went down. But all 3 drives are

> > > > > functional ... and I did the install to the raid itself. (mounted

> > > > > /dev/md1 and md2 before install) So nothing ever should have been out of

> > > > > sync.

> > > > > Any help as to why it's resynching or as to why it's already taken 24

> > > > > hours (and how to avoid this in the future) would be appreciated ...

> > > > >

> > > > > it's running gentoo

> > > > >

> > > > > Andrew Mathes

> > > > >

> > > >

> > > > --

> > > > "Streakers *repant* your end is in sight."

> > > >

> > >

> >

> > --

> > Hi, sorry I can't come to the phone right now. If you want to leave your name

> > and number, I'll phone you back. If you want to leave a fax, buy me a fax

> > machine.

> >


--
"Usenet is like a herd of performing elephants with diarrhea - massive,
difficult to redirect, awe-inspiring, entertaining, and a source of mind -
boggling amounts of excrement when you least expect it." - Gene Spafford,1992


More information about the Linux mailing list