Ignite Recovery Excessive I/O Errors

[Sean McConkey_3]

I have a 10.20 K-Class Ignite tape that fails after the initial recovery & first reboot. It starts to create the special files, hangs and then terminates with excessive I/O errors.

This NOT a H/W error on the box, as it happens using 3 different machines. And using 3 different Ignite tapes from the original machine.

Any ideas?

[Jerome Baron]

Hi,

Can you give us the exact error message and ignite version ?
Are you clone system or recovery the same system ?

Regards,
Jerome

[Sean McConkey_3]

The version they are using is B.3.6.82

The message is included below;

Boot
: disc(10/0.4.0;0)/stand/vmunix
4867116 + 569344 + 6122048 start 0x188868
vuseg=1438d000
inet_clts:ok inet_cots:ok Networking memory for fragment reassembly is restri
cted to 288223232 bytes
Logical volume 64, 0x3 configured as ROOT
Logical volume 64, 0x2 configured as SWAP
Logical volume 64, 0x2 configured as DUMP
Swap device table: (start & size given in 512-byte blocks)
entry 0 - major is 64, minor is 0x2; start = 0, size = 2097152
WARNING: Insufficient space on dump device to save full crashdump.
Only 1073741824 of 4026532864 bytes will be saved.
The device file /dev/rroot does not exist or is not correct.
Automatic checks of the root file system not performed.
Processor 1 did not start within expected time period.
Processor 2 did not start within expected time period.
Processor 3 did not start within expected time period.
Processor 4 did not start within expected time period.
Processor 5 did not start within expected time period.
Starting the STREAMS daemons.
9245XB HP-UX (B.10.20) #1: Sun Jun 9 06:31:19 PDT 1996

Memory Information:
physical page size = 4096 bytes, logical page size = 4096 bytes
Physical: 3932160 Kbytes, lockable: 2648816 Kbytes, available: 2719532 Kbyte
s

insf: Installing special files for sdisk instance 0 address 10/0.0.0
insf: Installing special files for sdisk instance 1 address 10/0.4.0
insf: Installing special files for sdisk instance 2 address 10/0.5.0
insf: Installing special files for sdisk instance 3 address 10/0.6.0
insf: Installing special files for sdisk instance 36 address 10/12/5.2.0
insf: Installing special files for stape instance 1 address 10/12/5.0.0
DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility has started receiving excessive
errors from the I/O subsystem. I/O error entries will be lost
until the cause of the excessive I/O logging is corrected.
If the DEMLOG daemon is not active, use the DIAGSYSTEM command
in SYSDIAG to start it.
If the DEMLOG daemon is active, use the LOGTOOL utility in SYSDIAG
to determine which I/O subsystem is logging excessive errors.

[Clemens van Everdingen]

Hi,

Is your SCSI chain terminated ? And with the correct terminator ?

C.

[Sean McConkey_3]

Yes... As I mentioned earlier, the same message on 3 different machines, which all work with other Ignite tapes.

The full scenario is that there are 2 K460 Class machines, both near identical.

One recovers fine on the K580's I am using whilst the other fails. What's also odd is that both are using the same version of Ignite & the same method ie. Recover to Ignite the box.

[Dietmar Konermann]

What do you mean with "terminates with excessive I/O errors"? Does it panic? Does it hang?

[Clemens van Everdingen]

Hi,

did you check with STM/MSTM/XSTM what subsystem is logging the IO error's ?

Memory/Harddisk/SCSI Controller ?

C.

[Sean McConkey_3]

It hangs. No panic, a reset is needed. I have managed to get the system into single user mode, by booting interactively, but this is fairly pointless because the Ignite process hasn't fully completed, as after it should complete this stage it would then reboot and bring the system up.

I can't run STM etc. as I can't get the system upto a point I can run STM. I am not even sure if it's installed as it's a clients system, not mine. The main problem is that the error message gives sod all away, other then making you think it's a H/W error, which I know it is not.

[Clemens van Everdingen]

Hi,

Do both machines have the same cpu's
i.e: PA risc 1.1 or 2

Have seen similar issues with incompatible hardware !

C.

[Bill Hassell]

What machine was used to create the Ignite tape? The more serious part of the messages is the failure to start the other processors and the fact that /dev/root is having problems. /dev/root is the emergency device file used when serious LVM problems exist and the system is trying desparately to get a root filesystem started.

Ignite tapes are NOT interchangeable between different product lines. For instance, a V-class Ignite is useless in restoring a D-class or N-class. Similalrly, a K-class Ignite tape won't be useful on a SuperDome. I would strongly suspectt that the original system that created the Ignite tape is radically different, especially in disk I/O (ie, SE SCSI vs. HV-FWD vs. fibre).

[Sean McConkey_3]

Bill

I thought this was part of the problem, but as to why it can't find /dev/root I am lost.

OK more information as I have it.

Original System

K460 3 x 180MHz (PA-8000)
1Gb Memory

Recovery Systems

K580 6 x 240MHz (PA-8200)
4Gb Memory

K580 4 x 240MHz (PA-8200)
2Gb Memory

Effectively there are 2 K460's one of which works fine on the recovery box, and both have very similar spec.

I guess there could be a difference in CPU version son the K460's. I will try and get this info from the client.

I do Ignite recoveries week in week out and have recovered D's to K's etc. K's to K's, not a problem like this before.

[Dietmar Konermann]

Maybe the kernel on your backed up K460 is missing some of the PA8200 enablement patches. Since the Ignite kernel boots w/o problem (1st INSTALL boot), I think the problem should disappear after installing a recent HW enablement bundle on the K460 before creating the tape.

Best regards...
Dietmar.

[Sean McConkey_3]

Dietmar... sounds very plausible.

I'll get the client to check the package/bundles on both boxes. Perhaps the one that works has it enabled.