public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Software RAID/Filesystem problems?
@ 2002-05-28 18:34 kwijibo
  0 siblings, 0 replies; 2+ messages in thread
From: kwijibo @ 2002-05-28 18:34 UTC (permalink / raw)
  To: linux-kernel

Hello,

I have run into a problem with software raid with large filesystems.
I am not sure if this is a software raid limitation or a filesystem
limitation.  I currently have two 3ware raid controllers that have
1.1TB (yes, terrabyte) of space on each.  I want to stripe the two
hardware raid controllers using software raid but when I try to make
the raid using the command mkraid /dev/md0 it pukes out this:

DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C if unsure!
handling MD device /dev/md0
analyzing super-block
disk 0: /dev/sdb1, 1120597978kB, raid superblock at 1120597888kB
disk 1: /dev/sdc1, 1120597978kB, raid superblock at 1120597888kB
raid0: looking at sdb1
raid0:   comparing sdb1(1120597888) with sdb1(1120597888)
raid0:   END
raid0:   ==> UNIQUE
raid0: 1 zones
raid0: looking at sdc1
raid0:   comparing sdc1(1120597888) with sdb1(1120597888)
raid0:   EQUAL
raid0: FINAL 1 zones
raid0: zone 0
raid0: checking sdb1 ... contained as device 0
  (1120597888) is smallest!.
raid0: checking sdc1 ... contained as device 1
raid0: zone->nb_dev: 2, size: -2053771520
raid0: current zone offset: 1120597888
raid0: done.
raid0 : md_size is -2053771520 blocks.
raid0 : conf->smallest->size is -2053771520 blocks.
raid0 : nb_zone is 1.
raid0 : Allocating 8 bytes for hash.


Notice the huge negative numbers -2053771520.  I have tried this with
a couple different block sizes but it doesn't help.  When I build a 
filesystem
on it, in this case I am trying reiserfs, it builds it but in the 
message log I
have these errors:

09:00: rw=0, want=134217729, limit=-2053771520
attempt to access beyond end of device
09:00: rw=0, want=134217730, limit=-2053771520
attempt to access beyond end of device
09:00: rw=0, want=134217731, limit=-2053771520
attempt to access beyond end of device

There are quite a few, I didn't post them all.  Then when I mount the fs
it only reports it as about a 93GB filesystem instead of the 2.2TB it should
be.  I did this same configuration in RAID 1(mirror) and it all worked 
well, no
errors and the reported filesystem size was correct.  So I guess my question
is: Am I hitting a kernel limit here or a filesystem limit, or is this 
just a plain
ol bug?  What is the current maximum filesystem size?  I need to I can try
ext3 but it always takes a bit to format.  This is on RedHat 7.3 with the
stock 2.4.18-4smp kernel.  I also tried it with the standard 2.4.18 
kernel with
the same results.   Any suggestions/information would be appreciated.  Let
me know if any more info would help, I usually forget info that is 
important.

Here is a copy of my /etc/raidtab (currently set up for mirroring, not 
stripping)

raiddev /dev/md0
        raid-level      1
        nr-raid-disks   2
        persistent-superblock   1
        chunk-size      4
        device  /dev/sdb1
        raid-disk       0
        device  /dev/sdc1
        raid-disk       1

Thanks,
Steven





^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Software RAID/Filesystem problems?
@ 2002-05-29  0:38 Neil Brown
  0 siblings, 0 replies; 2+ messages in thread
From: Neil Brown @ 2002-05-29  0:38 UTC (permalink / raw)
  To: kwijibo; +Cc: linux-kernel

On Tuesday May 28, kwijibo@zianet.com wrote:
> Hello,
> 
> I have run into a problem with software raid with large filesystems.
> I am not sure if this is a software raid limitation or a filesystem
> limitation.  I currently have two 3ware raid controllers that have
> 1.1TB (yes, terrabyte) of space on each.  I want to stripe the two
> hardware raid controllers using software raid but when I try to make
> the raid using the command mkraid /dev/md0 it pukes out this:

sorry.  No-Can-Do.  Not in 2.4 anyway.  Maybe in 2.6...

2 time 1.1TB > 2TB.

2TB is the most you can access with 32bits addressing for 512byte
sectors.

You are hitting an error at (1K) block 134217729 which is 1 block past
2TB.

If you make the sdb1 and sdc1 partitions less than 1TB you have a
better change of it working.

NeilBrown

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2002-05-29  0:38 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-05-28 18:34 Software RAID/Filesystem problems? kwijibo
  -- strict thread matches above, loose matches on Subject: below --
2002-05-29  0:38 Neil Brown

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox