| From: | "scott(dot)marlowe" <scott(dot)marlowe(at)ihs(dot)com> |
|---|---|
| To: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
| Cc: | Jonathan Bartlett <johnnyb(at)eskimo(dot)com>, Ernest E Vogelsinger <ernest(at)vogelsinger(dot)at>, <shridhar_daithankar(at)persistent(dot)co(dot)in>, <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: Linux supports hot-swappable hardware? [was Re: |
| Date: | 2003-06-18 17:06:31 |
| Message-ID: | Pine.LNX.4.33.0306181104570.4992-100000@css120.ihs.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Wed, 18 Jun 2003, Bruce Momjian wrote:
> scott.marlowe wrote:
> > I've found that while it's a little harder to hot swap individual disks in
> > linux using sw RAID, the ability to make the raid behave exactly as I want
> > is worth it. Having lost a RAID5 set to a hw controller that simply had
> > the cable to two drives come loose but refused to accept them back into
> > the RAID5 after that without formatting them first, I'm no longer as wild
> > about hw raid controllers as I once was.
>
> It seems that RAID controllers seem to be as likely a failure point as
> disk drives.
Actually, the LSI cards can be setup to each run a RAID0 and then RAID1
them together, and if one card fails, the other keeps running. I.e. they
can run two or more cards as though they were a single device. It's
pretty slick. I'm just not happy with they way they behave when certain
things happen, like my story above.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2003-06-18 17:07:44 | Re: PostgreSQL calibration |
| Previous Message | scott.marlowe | 2003-06-18 17:04:25 | Re: A creepy story about dates. How to prevent it? |