Wurzel

iTrader: (1)

I need some help here as I am going insane.

I have a V240 with 4 x 72gb disks, I have mirrored disk 0 and disk 1 successfully and these work perfectly, the machine can boot from either disk.

I then mirrored disk 2 and disk 3 successfully.

I added the /dev/md/rdsk/d6 blah blah blah to vfstab so that it mounts automatically on a reboot.

My problem is that when the system reboots I get an error saying it can not read DISK BLOCK 143287232: Invalid argument

Then unexpected inconsitency and to FSCK it.

The system stops and wants me to login to maintenence mode and then press CTRL C but when I do this the machine ignore CTRL C and carries on booting. When I do df -k on the machine the offending d6 is mounted correctly and I can write to it.

I tried setting the mount at boot flag to no in vfstab and sure enough it doesn't mount and the machine boots normally, if I then try and mount the d6 disks manually it tells me that it is already mounted. However when I cd to the mount point there is nowt there. If I try to unmount it it says that it is busy and can't be unmounted.

If I try to fsck d6 it screws up.

Also if I try to recreate the mirror from scratch even after blowing away the disks and newfsing them it tells me that the pairing already exists, inspite of me blowing away the database slice 7 on each disk and recreating the vtoc on slice 2.

Can anyone shed any light on what could be causing this boot problem.

Cheers

Steve

stevem2k

CTRL-D after maintenence....

is d6 a whole disc or slice ?

Wurzel

iTrader: (1)

it is a 67gb slice.

I made c1t2d0s0 the slice I want mounting and I made

c1t2d0s7 the SVM replica slice.

I thought about using slice 2 as technically need the whole disk as one partition but I needed to make the SVM replica in order to mirror the disks so decided on slice 0 and 7.

Does this make sense?

The thing is want it to mount normally I do not want to have to CTRL-C or CTRL-D anything.

stevem2k

what does metadb -i say ?

I think your state databases aren't right for d6 which is why it boots normally with d6 disabled.

Wurzel

iTrader: (1)

Ok now I am really confused?

if I do metastat d6 it tells me that the size is

141266880 blocks (67 GB)

but the error on boot is saying it can not read DISK BLOCK 143287232

Why is it even trying to read that block?

Wurzel

iTrader: (1)

root@tcsun29: metadb -i
flags first blk block count
a m p luo 16 8192 /dev/dsk/c1t0d0s7
a p luo 8208 8192 /dev/dsk/c1t0d0s7
a p luo 16 8192 /dev/dsk/c1t1d0s7
a p luo 8208 8192 /dev/dsk/c1t1d0s7
F M p 16 unknown /dev/dsk/c1t2d0s7
a p luo 8208 8192 /dev/dsk/c1t2d0s7
F M p 16 unknown /dev/dsk/c1t3d0s7
a p luo 8208 8192 /dev/dsk/c1t3d0s7

r - replica does not have device relocation information
o - replica active prior to last mddb configuration change
u - replica is up to date
l - locator for this replica was read successfully
c - replica's location was in /etc/lvm/mddb.cf
p - replica's location was patched in kernel
m - replica is master, this is replica selected as input
W - replica has device write errors
a - replica is active, commits are occurring to this replica
M - replica had problem with master blocks
D - replica had problem with data blocks
F - replica had format problems
S - replica is too small to hold current data base
R - replica had device read errors

it appears the database on disk 2 and 3 are corrupt, so how can I recreate them?

stevem2k

There's your problem....

Wurzel

iTrader: (1)

yeah but what do I do to rectify it?

if I run this again it tells me c1t2d0s7 already has a metadevice database replica.

# metadb -a -f -c 2 c1t2d0s7
# metadb -a -c 2 c1t3d0s7

stevem2k

to clarify :

you have a mirror of c1t0d0 and c1t1d0 .

you have *added* a second mirror of c1t2d0 and c1t3d0

so 2x raid-1 yes?

Wurzel

iTrader: (1)

Correct!

stevem2k

** I assume there is no data on the t2/t3 mirror ?

I don't expect this to be a problem but how big is s7 ?

if you do
What does it look like ?

Wurzel

iTrader: (1)

it looks like this

Sorry s7 is 1gb

root@tcsun29: metadb
flags first blk block count
a m p luo 16 8192 /dev/dsk/c1t0d0s7
a p luo 8208 8192 /dev/dsk/c1t0d0s7
a p luo 16 8192 /dev/dsk/c1t1d0s7
a p luo 8208 8192 /dev/dsk/c1t1d0s7
a u 16 8192 /dev/dsk/c1t2d0s7
a u 8208 8192 /dev/dsk/c1t2d0s7
a u 16400 8192 /dev/dsk/c1t2d0s7
a u 16 8192 /dev/dsk/c1t3d0s7
a u 8208 8192 /dev/dsk/c1t3d0s7
a u 16400 8192 /dev/dsk/c1t3d0s7
root@tcsun29:

Wurzel

iTrader: (1)

Ok that has resolved the BLOCK issue but the error has now changed to

Can't roll the log for /dev/md/rdsk/d6

It wants me to fsck the filesystem

then after it resumes it says

resuming mountall
mount I/O error
mount: cannot mount /dev/md/dsk/d6

so should I detach the mirror and then reattach it so that it syncs again?

stevem2k

yes, break the mirror , fsck the file system and reattach.

Wurzel

iTrader: (1)

ok will do and then will report back

stevem2k

rate card is 900/day, I'll pm you with where to send the Purchase Order

Wurzel

iTrader: (1)

ok

anyway I broke the mirror and tried to fsck the slice and it failed.

It told me FSCK was running in yes mode and if I wanted to do this I would need to run a certain command blah blah

so I did and it told me it could not read DISK Block 143287232 I/O error

then it said it could not read sectors 143287232 143287233 143287234 143287235

stevem2k

<sigh>

format for a Seagate 73G says it has

backup wu 0 - 14086 68.35GB (14087/0/0) 143349312

Which would put your 'errant' blocks in the last 62080 - or s7

..so why is d6 trying to access this ?

When you did the formatting did you label the disc ????

Wurzel

iTrader: (1)

Yes I labeled c1t2 as js1 and c1t3 as js2.

Sorry for the late reply I am busy documenting the mirror process.

root@jumpstart: format
Searching for disks...done


AVAILABLE DISK SELECTIONS:
0. c1t0d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1c,600000/scsi@2/sd@0,0
1. c1t1d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1c,600000/scsi@2/sd@1,0
2. c1t2d0 <SEAGATE-ST373307LC-0007 cyl 49780 alt 2 hd 4 sec 720>
/pci@1c,600000/scsi@2/sd@2,0
3. c1t3d0 <SEAGATE-ST373307LC-0007 cyl 49780 alt 2 hd 4 sec 720>
/pci@1c,600000/scsi@2/sd@3,0
Specify disk (enter its number):

Wurzel

iTrader: (1)

Ok update time, I could not get the d6 slice to mount or fsck properly so I configured my jumpstart server to create all the partitions I needed on the 2 disks I was going to mirror so c1t0d0 and c1t2d0, and reinstalled the machine.

Then I mirrored c1t2d0 to c1t3d0 and it all worked perfectly. Then I mirrored the root disk c1t0d0 to c1t1d0 and that also worked perfectly, the machine is now all mirrored up and ready for me to start to build my new jumpstart server. I have no idea what was wrong with the d6 slice when I manually created it, maybe it got it's knockers in a twist about summat. Anyway it now all works as I wanted it to.

Cheers for your help Steve. I had no idea how to delete the metadb until you gave me the command. I have now documented the process so that even a manager can understand it

stevem2k

If you said you were doing it via jumpstart I would have sent you one of my jet templates that does all the mirroring for you

Glad it's all ok now anyway ...


S

Wurzel

iTrader: (1)

How about you send it to me anyway as I am always keen an learning new stuff.

Cheers

Steve

stevem2k

couple of examples inbound

Wurzel

iTrader: (1)

Cheers.