11.04 RAID (re)install issues

[maclenin]

While my RAID arrays complete their resync, I thought I'd jot this quick line to see if anyone had any thoughts as to why I am seeing the following sfdisk ERROR with regard to the status of my RAID arrays?

sfdisk: ERROR: sector 0 does not have an MSDOS signature
/dev/md0: unrecognised partition table type
No partitions found
I have just completed a "clean" install of 11.04 (from which I am composing this note) - replacing an existing 3-disk (sda, sdb, sdc) RAID/LVM set-up (8.10).

disk1 - sda 250 gb
disk2 - sdb 1 tb
disk3 - sdc 1 tb

Data is all backed-up.

I chose not to keep the olde RAID/LVM configuration, as there were changes I wanted to implement....

I'd like to resolve the RAID ERRORs before configuring LVM.

My current(ly resync-ing) RAID configuration:

md0
sda1 RAID1 /boot
sdb1 RAID1 /boot
sdc1 RAID1 /boot

md1
sda3 RAID5 /
sdb3 RAID5 /
sdc3 RAID5 /

md2
sdb4 RAID1 /shared
sdc4 RAID1 /shared

I also have (non-RAID) swap on:

sda2
sdb2
sdc2

$ cat /proc/mdstat

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md2 : active raid1 sdb4[0] sdc4[1]
731546488 blocks super 1.2 [2/2] [UU]
[==>..................] resync = 10.2% (74791168/731546488) finish=192.0min speed=56984K/sec

md0 : active raid1 sdc1[2] sdb1[1] sda1[0]
487412 blocks super 1.2 [3/3] [UUU]

md1 : active raid5 sdb3[1] sda3[0] sdc3[2]
485460992 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU]

unused devices: <none>
$

$ sudo sfdisk -l -uS

Disk /dev/sda: 30522 cylinders, 255 heads, 63 sectors/track
Units = sectors of 512 bytes, counting from 0

Device Boot Start End #sectors Id System
/dev/sda1 * 2048 976895 974848 fd Linux RAID autodetect
/dev/sda2 976896 4882431 3905536 82 Linux swap / Solaris
/dev/sda3 4882432 490346495 485464064 fd Linux RAID autodetect
/dev/sda4 0 - 0 0 Empty <-----curious?

Disk /dev/sdc: 121601 cylinders, 255 heads, 63 sectors/track
Units = sectors of 512 bytes, counting from 0

Device Boot Start End #sectors Id System
/dev/sdc1 * 2048 976895 974848 fd Linux RAID autodetect
/dev/sdc2 976896 4882431 3905536 82 Linux swap / Solaris
/dev/sdc3 4882432 490428415 485545984 fd Linux RAID autodetect
/dev/sdc4 490428416 1953523711 1463095296 fd Linux RAID autodetect

Disk /dev/sdb: 121601 cylinders, 255 heads, 63 sectors/track
Units = sectors of 512 bytes, counting from 0

Device Boot Start End #sectors Id System
/dev/sdb1 * 2048 976895 974848 fd Linux RAID autodetect
/dev/sdb2 976896 4882431 3905536 82 Linux swap / Solaris
/dev/sdb3 4882432 490428415 485545984 fd Linux RAID autodetect
/dev/sdb4 490428416 1953523711 1463095296 fd Linux RAID autodetect

Disk /dev/md1: 121365248 cylinders, 2 heads, 4 sectors/track

sfdisk: ERROR: sector 0 does not have an MSDOS signature
/dev/md1: unrecognised partition table type
No partitions found

Disk /dev/md0: 121853 cylinders, 2 heads, 4 sectors/track

sfdisk: ERROR: sector 0 does not have an MSDOS signature
/dev/md0: unrecognised partition table type
No partitions found

Disk /dev/md2: 182886622 cylinders, 2 heads, 4 sectors/track

sfdisk: ERROR: sector 0 does not have an MSDOS signature
/dev/md2: unrecognised partition table type
No partitions found
$

Thanks for any insight!

[maclenin]

Just a quick follow-up - the arrays have completed their resync and the ERRORs remain the same....

[YesWeCan]

Looks normal to me.
sfdisk cannot find an MSDOS signature on your "virtual device" md0. An mdadm device does not have an MBR. If it said this on one of your real devices, like sda, then that would be more interesting.

The "curious" thing is simply that every MBR has 4 partition table slots. Your sda is only using 3 of them.

[maclenin]

Thanks, YesWeCan, for the comments. Perhaps I should clarify a wee bit....

I have installed - or tried to (re)install my RAID arrays over top an already existing RAID/LVM set-up (which was working beautifully in 8.10).... The reason I upgraded was that 8.10 was getting old and I wanted to change some of the RAID configs. There were other features that needed a re-fresh, as well...so, it wasn't, simply, a yen for the bleeding edge....

With regard to sda - and the MBR's 4 partition table slots....

My drives / RAID arrays are as follows:

Drive 1 - 250 gb
sda1 - boot - md0 [RAID1]
sda2 - swap
sda3 - root - md1 [RAID5]
sda4 - ???

Drive 2 - 1 tb
sdb1 - boot - md0 [RAID1]
sdb2 - swap
sdb3 - root - md1 [RAID5]
sdb4 - shared - md2 [RAID1a]

Drive 3 - 1 tb
sdc1 - boot - md0 [RAID1]
sdc2 - swap
sdc3 - root - md1 [RAID5]
sdc4 - shared - md2 [RAID1a]

My RAID setup was designed to distribute drive space evenly across the drives/RAID arrays. The smallest drive (1) was "replicated" across 3 drives - taking 250 gb from each. The largest two drives split what was left (750gb each) to create /shared....

drive1(sda)=md0 + swap + md1 = 250 gb
drive2(sbd)=md0 + swap + md1 = 250 gb + md2 (750 gb) = 1000 gb
drive3(sdc)=md0 + swap + md1 = 250 gb + md2 (750 gb) = 1000 gb

And with further regard to sda and the ghostly number 4 - take a peek at this page (http://www.howtoforge.com/software-raid1-grub-boot-fedora-8-p4) - about 2/3 of the way down, where you'll see:


[root@server1 ~]# sfdisk -d /dev/sda | sfdisk /dev/sdb
Checking that no-one is using this disk right now ...
OK

Disk /dev/sdb: 652 cylinders, 255 heads, 63 sectors/track

sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/sdb: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = sectors of 512 bytes, counting from 0

Device Boot Start End #sectors Id System
/dev/sdb1 * 63 208844 208782 fd Linux raid autodetect
/dev/sdb2 208845 1237004 1028160 fd Linux raid autodetect
/dev/sdb3 1237005 10474379 9237375 fd Linux raid autodetect
/dev/sdb4 0 - 0 0 Empty <-----curiouser?
Successfully wrote the new partition table

Re-reading the partition table ...

If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
[root@server1 ~]#

I hope this all helps with the untangling....

Anything you'd suggest I try - to kick the tires?

[YesWeCan]

Sorry, I can't help. I do not understand what the problem is.

[maclenin]

No worries YesWeCan (I still do think so....) - thanks for the leads (MBR, perhaps?) - I'll keep looking - and report back if I find anything.... In the meantime - all are welcome to toss any pearls my way!

[maclenin]

I have - for the time being - ambled into a "standard" (non_RAID / non_LVM) install 'cause I have work I need to do, but I am still curious (for near-future reference) to see whether anyone has any ideas as to why the sfdisk ERRORs might have crept into my earlier RAID setup....

Thanks for any comments!