Heinz Christian wrote:
> > > From: Mike Bilow[SMTP:[EMAIL PROTECTED]]
> > > The aic7xxx driver really does not like to share an IRQ.
> Have you enabled the MP-APIC-Support?
>
> We have several machines each with 5 or more AIC7xxx channels with APIC
> en
In theory yes, in practice no. The aic7xxx is a high-performance,
time-critical driver. If you make it share an IRQ, then it loses a
certain amount of control to its brother drivers. This is especially
annoying with motherboards which have both embedded aic7xx and Ethernet
hardware on board, as
I think it means you tried to run fsck across an area being resynced.
-- Mike
On Wed, 12 Apr 2000 [EMAIL PROTECTED] wrote:
> Hello!
>
> This is what I got in the dmesg after my machine with raid5 crashed again:
>
> set_blocksize: b_count 1, dev md(9,0), block 245!
> set_blocksize: b_count 1,
The aic7xxx driver really does not like to share an IRQ.
-- Mike
On Wed, 12 Apr 2000, Morten Bøgeskov wrote:
> I've done this and added dvd-ide, reiser & supermount. But this is not
> really working 8). Does anybody else have the problem that an insmod
> aic7xxx (which is my boot controller) r
On Tue, 11 Apr 2000, Darren Nickerson wrote:
> > On Tue, 11 Apr 2000, "Lance" == D. Lance Robinson wrote:
>
> Lance> So, if the md driver doesn't fail a drive that is because the lower
> Lance> levels have taken care of all the nitty details and have supposedly
> Lance> performed the r
On Sat, 8 Apr 2000, Edward Schernau wrote:
> Mike Bilow wrote:
> >
> > Drives use by far more instantaneous current to spin up then they
> need to
> > run in active idle mode. A big, modern IDE drive can consume
> about 2.0A
>
> [Excellent analysis trimmed]
Not only is it perfectly reproducible on your machine, it was on mine,
too! Get the new kernel patch from Ingo Molnar:
http://people.redhat.com/mingo/raid-patches/raid-2.2.14-B1
and the accompanying tools:
http://people.redhat.com/mingo/raid-patches/raidtools-dangerous-0.90-20
Do you see anything in the log? SCSI parity error, for example?
With three identical drives on the same controller, that rules out a lot
of issues that could arise. If you are seeing the same drive being marked
bad over and over, and it takes days for that to happen, I would look at
(1) termina
I have to agree with Leonard (big surprise)... we have had two large
server machines in continuous service for a while with the Mylex RAID
controllers, busy 24 hours a day, and we have never seen anything that
approaches the problems you are describing. My bet is also hardware.
-- Mike
On Fri,
Drives use by far more instantaneous current to spin up then they need to
run in active idle mode. A big, modern IDE drive can consume about 2.0A
to spin up, but probably no more than 0.5A to sit in active idle. (Note
that the "." is a decimal point; I'm in the USA.)
For example, I looked up th
Why on earth would you even bother with RAID on a machine which can
experience 30 minutes of idle time? We only use RAID for machines which
run web servers, mail servers, name servers, and things like this. Such
machines here never exeperience even a solid minute of idle time.
-- Mike
On Thu,
I am not sure if anyone actually answered your question. Here is a
lilo.conf that worked for me:
boot=/dev/md0
root=/dev/md0
install=/boot/boot.b
map=/boot/map
delay=20
vga=normal
append="panic=120"
lba32
default=Linux
image=/boot/vmlinuz
label=Linux
read-only
image=/boot/vmlinuz
I think you want: http://www.linux-ha.org/
In general, it is better to have two independent servers that replicate
data back and forth than to try to put multiple initiators on a common
storage bus. SCSI does allow this in theory, but as far as I know Linux
does implement the necessary commands
Did you leave the "update" daemon running?
-- Mike
On Mon, 3 Apr 2000, Paramasivam Kartik wrote:
> Hi everybody,
>
> this is a question about the linux source code.
>
> i want to test my modified raid code for reads only.
> Once after every couple of read requests,
> a write request is auto
In general, no. There is a master partition table which occupies the
first sector of the device and which could therefore be copied. However,
this master partition table may link to other "extended" partition tables
which can be anywhere on the drive.
I have not tested this, but I assume someth
I have a related idea. If all of the information in a raidtab can be
derived from a persistent superblock, there should be a tool that allows
generating an actual raidtab file by reading a persistent superblock.
This would be especially useful for dianostic procedures.
Also, is there any downsid
On Sat, 1 Apr 2000, Mike Bilow wrote:
> On Fri, 31 Mar 2000, Michael wrote:
> > hmmm. the remirroring code is not very smart... as I recall it
> > does the remirroring in order .. i.e. md0, md1, etc... This would
> > imply that if you have a power fail or other c
http://linux-ha.org/
-- Mike
On Sun, 2 Apr 2000, octave klaba wrote:
> Hi,
> We are looking for the hardware/software solutions to make
> clustering on Linux. We want to have 2 servers with 1 raid
> storage or a solution to write on 2 servers in the same time.
On Fri, 31 Mar 2000, Michael wrote:
> > /dev/md0/dev/hda1 + /dev/hdc129.7 GBRAID-1
> > /dev/md1/dev/hda2 + /dev/hdc2 0.3 GBRAID-1
> >
> > We use /dev/md0 for the root fs and /dev/md1 for swap. Why?
> > Because it takes about 90 minutes to remirror /dev/md0 and only
> >
You are missing that nasty "--absolute-paths" ("-P") switch on tar to
preserve the leading slash. It may or may not be what you want. In
general, a better approach is to tell tar to make a particular directory
the current directory before executing, using the "--directory" ("-C")
switch. For ex
On Fri, 31 Mar 2000, Michael wrote:
> > /dev/md0/dev/hda1 + /dev/hdc129.7 GBRAID-1
> > /dev/md1/dev/hda2 + /dev/hdc2 0.3 GBRAID-1
> >
> > We use /dev/md0 for the root fs and /dev/md1 for swap. Why?
> > Because it takes about 90 minutes to remirror /dev/md0 and only
> >
Why are you backing up? What if your machine catches fire? What if your
office catches fire? What if someone steals your machine? What if your
power supply sends mains voltage to every device in the case?
-- Mike
On Fri, 31 Mar 2000, Jeff Hill wrote:
> I know this is off-topic, but since i
Probably because the installer does not handle it. I have been working on
the Debian end of the same issue, and the problems pretty much amount to
this, based on the current potato snapshot:
1. The kernel has to be patched with the newest RAID v0.90 code and
rebuilt, and then put onto the first
I think there is rarely a valid reason to split a single disk system into
multiple small partitions. In fact, the Multi-Disk HOWTO agrees:
In fact, for single physical drives this scheme offers very little
gains at all, other than making file growth monitoring easier
(usi
I would think that a block device is a block device from the point of view
of software like this, so I cannot see any reason to expect a block device
created by software RAID to look any different to it. That said, I have
never tried it, so take my advice with a grain of salt.
-- Mike
On Wed,
AIL PROTECTED] wrote:
> Mike Bilow <[EMAIL PROTECTED]> wrote:
> >...snip..
> >However, if you have IO-APIC support enabled in the CMOS setup
> >and in the Linux kernel, there should be no need for assignment
> >of shared interrupts.
> >...snip...
>
>
ture of IO-APIC.
Hi, Neighbor!
-- Mike
--
---
Bilow Computer Science, Inc. | http://www.bilow.com/ | Michael S. Bilow
Cranston, RI 02920-5554, USA | [EMAIL PROTECTED]
I am no RAID expert, but I know a SCSI bus hang when I see one.
A suggestions which I am sure everyone will make:
1. Try the 2.2.14 kernel with the newer patch:
http://people.redhat.com/mingo/raid-patches/raid-2.2.14-B1
And a few of my own:
2. Try booting with kernel argument "aic7xxx
Danilo Godec wrote:
>
> On Tue, 28 Mar 2000, Mike Bilow wrote:
>
> > I think you have an electrical issue.
>
> I feared that, but what should I do? It's all LVD and all pre-installed by
> Intel... except disks, of course.
Call for warranty service, I would think
On Tue, 28 Mar 2000, octave klaba wrote:
> > Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
> > > > 1° 2000 9212 96.4 20354 13.9 4411 3.6 3822 35.4 22180 8.0 85.6 0.7
> > > > 2° 2000 1727 22.4 2095 5.5 1381 34.9 3070 98.6 4320 97.8 74.8 7.3
> > > > 3° 2
APIC-level aic7xxx, aic7xxx
* * *
> 1400-14be : aic7xxx
> 1800-18be : aic7xxx
Are you really running two separate aic7xxx controllers? Do they have the
same firmware revision?
-- Mike
--
---
Bilow Computer S
On Tue, 28 Mar 2000, Jakob Østergaard wrote:
> On Tue, 28 Mar 2000, octave klaba wrote:
>
> > Hi,
> >
> > a test between single ide <-> raid-1 soft ide <-> raid-1 soft scsi-2
> >
> > 1° PIII600/128RAM/1XIDE20.5
> > 2° PIII600/128RAM/2XIDE20.5 raid-1 soft
> > 3° PIII500/256/29
The aic7xxx driver is definitely unstable in my experience on SMP when
shared interrupts are used. It will usually hang on boot in this case.
This is an especially annoying problem because many SMP motherboards will
insist on assigning interrupts automatically; Intel and Supermicro are
particul
I have not really done this yet, but
if [ ! -e /proc/mdstat ] || \
[ `grep -ci resync /proc/mdstat` -eq 0 ] ; then
swapon -a
fi
seems like a reasonable approach. Martin Bene posted a slightly different
variant which waits until remirroring is done and then st
t free to experiment.
-- Mike
On Sat, 25 Mar 2000, Mike Bilow wrote:
> Unfortunately, the boot floppy fails out with a kernel panic about being
> unable to mount the root filesystem. The ssyslinux.cfg file does
> correctly specify "root=/dev/md0" and the VFS message reports tha
Starting with the Debian Potato Rescue/Root disks, I replaced the kernel
with a custom build of 2.2.14 with Ingo's patches. I booted from this
modified Debian install set and (using fdisk) manually created two type
0xFD partitions, /dev/hda1 and /dev/hdc2. I then ran mkraid and linked
these two
What is the consensus opinion about whether swapping onto RAID is a good
idea? I understand from various people who should know that swapping onto
RAID can be unstable when actually remirroring. Does anyone have personal
experience with this issue?
I am building a dual-IDE system with two ident
I am extremely inexperienced with software RAID, so please don't flame me
if this message comes arcross as evidence of outright idiocy. I prefer to
think the line between idiocy and adventurousness is just very thin.
I built a custom kernel using the 2.2.14 source and applying the most
recent pa
38 matches
Mail list logo