dd on root disk

[Karthik S S]

Hi,

I have A500 system with 2 * 18 GB HDDs. Disk1 has the OS and Disk2 is not used. I am often getting the following error in the syslog,

-------------------------------
Jan 19 10:01:08 hpux4 vmunix: LVM: vg[0]: pvnum=0 (dev_t=0x1f01f000) is POWERFAILED
Jan 19 10:01:08 hpux4 vmunix: SCSI: Unexpected Disconnect -- lbolt: 22865735, dev: 1f01f000, io_id: 10210d8
Jan 19 10:01:08 hpux4 vmunix: SCSI: Unexpected Disconnect -- lbolt: 22865735, dev: 1f01f000, io_id: 10210d9
Jan 19 10:01:08 hpux4 vmunix: LVM: Recovered Path (device 0x1f01f000) to PV 0 in VG 0.
Jan 19 10:01:08 hpux4 vmunix: LVM: Restored PV 0 to VG 0.
-------------------------------

I assume there is some problem with the Disk1. So I am planning to install the OS on the Disk2. Instead of freshly installing and configuring the system, can I do a "dd" of disk1 on to disk2?. This obviously has to be done when the system is booted with disk1. What are all the things need to be considered for doing this?. What happens to the open files? Is there any other better method for replicating the disks (other than MirrorDisk-UX and Ignite)

Pl. help,
Thanks,
Karthik S S

[Leif Halvarsson_2 · Accepted solution]

Hi,
Possible alternatives if you don't want to use Ignite or MirrorDisk.
- If you use Omniback/DataProtector, do a rawdisk backup of disk 1 and restore to disk 2.
- Create a bootable disk as the examples in lvlnboot and copy the filesystems to the new disk (use cpio). I have done it a number of times, it is rather complex but may be necessary if the two disks are different.

[Stefan Farrelly]

You can do a dd, but do it when the system is booted in single user mode so not much running. It works fine.

Your LVM errors show the device seems to be powering down then up again, a sure sign its on its way to disk heaven.

[Karthik S S]

But is it not possible with dd command??

-Karthik S S

[T G Manikandan]

you cannot do anything until the problem gets resolved because the SCSI hard disk times out.Even you do a dd it will hang in the middle of the process.

Just check for the scsi patches and scsi termination.

Just do diskinfo on that disk.

[Sunil Sharma_1]

It will work Karthik .
don't worry

Sunil

[T G Manikandan]

I think with the proper scsi patches,if the drive is not bad with the diskinfo output.

increase the timeout on that disk.It should resolve things

#pvchange -t 180 /dev/dsk/....

[Stefan Farrelly]

It is possible with the dd command. Try it!

[Bill Hassell]

You can use dd but to avoid problems, perform the task in signle user mode. That eliminates any changing data during the dd copy. You'll need to mount /usr to gain access to the dd executable. Once you've finished copying (be sure to use a large blocksize like bs=32k or it will take hours), you'll have to shutdown and power off the box so you can exchange the SCSI addresses on the disks. The dd copy has the physical address of the boot disk so it must match the hardware. There is no way to change this during bootup.

[Karthik S S]

Hi Manikandan,

Where do I get the SCSI patches for HP-UX 11.11?

Thanks,
Karthik S S

[Michael Tully]

To get to the patch start here:

http://www1.itrc.hp.com/service/patch/pdbMainPage.do

select find patches:

Type in SCSI where is asks for the keywork and have 11.11 as your selection version.

[T G Manikandan]

PHCO_23651 fsck_vxfs(1M) cumulative patch
PHCO_25902 cumulative SAM/ObAM patch
PHCO_26017 user/group(add/mod/del)(1M) cumulative patch
PHCO_27369 LVM commands cumulative patch
PHKL_18543 PM/VM/UFS/async/scsi/io/DMAPI/JFS/perf patch
PHKL_20016 2nd CPU not recognized in G70/H70/I70
PHKL_24027 VxFS 3.1 cumulative patch
PHKL_24187 ioscan performance gain for SCSI Subsystem
PHKL_25475 PM cumulative patch
PHKL_27003 SCSI IO Subsystem Cumulative Patch
PHKL_27178 LVM Cumulative Patch w/Performance Upgrades
PHKL_27364 Probe,IDDS,PM,VM,PA-8700,AIO,T600,FS,Hang

the above patches are for 11.00.
find their 11.11 eqv.

The best way would be to apply the recent patch bundles on the OS.

[Karthik S S]

Hi,

Thanks for that. I did a DD on the other disk but still the problem persisits. I am getting the following error. Is this a SCSI Bus problem?? I mean is it a Hardware or software problem??

---------------------

SCSI: Resetting SCSI -- lbolt: 465191, bus: 1
SCSI: Reset detected -- lbolt: 465191, bus: 1
SCSI: Unexpected Disconnect -- lbolt: 477011, dev: 1f01f000, io_id: 1004ac5
SCSI: Unexpected Disconnect -- lbolt: 541633, dev: 1f01f000, io_id: 1004c17
SCSI: Unexpected Disconnect -- lbolt: 541811, dev: 1f01f000, io_id: 1004c1d
SCSI: Unexpected Disconnect -- lbolt: 1119151, dev: 1f01f000, io_id: 10056a3
SCSI: Third party detected bus hang -- lbolt: 1409791, bus: 1
lbp->state: 60
lbp->offset: f0
lbp->uPhysScript: f8040000
From most recent interrupt:
ISTAT: 01, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000010
lsp: 0000000000000000
lbp->owner: 000000004108df00
bp->b_dev: 1f01f000
scb->io_id: 1005bf2
scb->cdb: 2a 00 00 4a ef f4 00 00 04 00
lbolt_at_timeout: 1409691, lbolt_at_start: 1408191
lsp->state: 85
SCSI: Resetting SCSI -- lbolt: 1409891, bus: 1
SCSI: Reset detected -- lbolt: 1409891, bus: 1
SCSI: Unexpected Disconnect -- lbolt: 1539450, dev: 1f01f000, io_id: 1005eba
SCSI: Third party detected bus hang -- lbolt: 1948391, bus: 1
lbp->state: 60
lbp->offset: 80
lbp->uPhysScript: f8040000
From most recent interrupt:
ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000001
lsp: 0000000000000000
lbp->owner: 00000000425c8900
bp->b_dev: 1f01f000
scb->io_id: 1006642
scb->cdb: 2a 00 00 43 6f c0 00 00 10 00
lbolt_at_timeout: 1948291, lbolt_at_start: 1946791
lsp->state: 85
SCSI: Resetting SCSI -- lbolt: 1948491, bus: 1
SCSI: Reset detected -- lbolt: 1948491, bus: 1
SCSI: Third party detected bus hang -- lbolt: 2134291, bus: 1
lbp->state: 60
lbp->offset: 80
lbp->uPhysScript: f8040000
From most recent interrupt:
ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000001
lsp: 0000000000000000
lbp->owner: 00000000425c8300
bp->b_dev: 1f01f000
scb->io_id: 10069d1
scb->cdb: 2a 00 00 5e fc 28 00 00 04 00
lbolt_at_timeout: 2134191, lbolt_at_start: 2132691
lsp->state: 85
SCSI: Resetting SCSI -- lbolt: 2134391, bus: 1
SCSI: Reset detected -- lbolt: 2134391, bus: 1
SCSI: Unexpected Disconnect -- lbolt: 2179907, dev: 1f01f000, io_id: 1006adb
SCSI: Unexpected Disconnect -- lbolt: 2220411, dev: 1f01f000, io_id: 1006bd9
SCSI: Unexpected Disconnect -- lbolt: 2601111, dev: 1f01f000, io_id: 10072bd
SCSI: Unexpected Disconnect -- lbolt: 3620127, dev: 1f01f000, io_id: 1008585
SCSI: Unexpected Disconnect -- lbolt: 3839934, dev: 1f01f000, io_id: 1008d15
SCSI: Third party detected bus hang -- lbolt: 4182891, bus: 1
lbp->state: 4060
lbp->offset: 80
lbp->uPhysScript: f8040000
From most recent interrupt:
ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000001
lsp: 0000000000000000
lbp->owner: 0000000041eadb00
bp->b_dev: 1f01f000
scb->io_id: 1009acf
scb->cdb: 2a 00 00 29 96 f0 00 00 04 00
lbolt_at_timeout: 4182791, lbolt_at_start: 4181291
lsp->state: 85
SCSI: Resetting SCSI -- lbolt: 4182991, bus: 1
SCSI: Reset detected -- lbolt: 4182991, bus: 1
SCSI: Third party detected bus hang -- lbolt: 4657891, bus: 1
lbp->state: 4060
lbp->offset: 80
lbp->uPhysScript: f8040000
From most recent interrupt:
ISTAT: 09, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000001
lsp: 0000000000000000
lbp->owner: 0000000041ead900
bp->b_dev: 1f01f000
scb->io_id: 100a408
scb->cdb: 2a 00 00 2a ac 7c 00 00 04 00
lbolt_at_timeout: 4657791, lbolt_at_start: 4656291
lsp->state: 85
SCSI: Resetting SCSI -- lbolt: 4657991, bus: 1
SCSI: Reset detected -- lbolt: 4657991, bus: 1
-----------------------------

Thanks,
Karthik S S

[T G Manikandan]

1.//check the scsi termination// & scsi cabling

2.Increase the timeout on that disk using
pvchange

[Karthik S S]

But these are 2 internal disks on A500 system and no other external boxes are connected. So I do not suspect the termination/cabling. Also I removed the disks once and put it back. I have changed the timeout vaule to 180 for the physical disk.

Thanks
Karthik S S

[Karthik S S]

I loaded the following SCSI patches,

PHCO_23510 PHKL_27498 PHKL_29365 PHKL_29685 PHCO_29098 PHKL_29248 PHKL_29468

But still the system reports the same lbolt messages :-(

-Karthik S S

[T G Manikandan]

Just verify the disk using

#dd if=/dev/rdsk/ of=/dev/null bs=1024k

check that there are no I/O errors.

[Karthik S S]

Infact when I did a dd to copy the data to the other disk I did not get any IO error.

Thanks,
Karthik S S

[Michael Tully]

This may sound like a dumb idea ... but did the copy of the data take an unusually long time ? Just in case (and you didn't hear this from me), but I would use mirrordisk/UX to copy the data temperarily to the second disk and see what happens. (as long as you have another license somewhere you can use) There is certainly something wrong somewhere, and it is still likely to be a disk or SCSI problem. If it ends up being a disk, then at least you can replace the disk without too many dramas ...

[Karthik S S]

Hi Michael,

dd of the root disk to the other internal disk (both are of 17GB) took about an hour with the block size of 1024k. Then I swapped the disks between the two slots of A500 and realized that the lbolt messages are always referring to the disk in the upper slot. Hence, I am reinstalling the OS on the lower slot disk. Yes it is getting dramatic only :-))

Thanks,
Karthik S S