Unsolved

This post is more than 5 years old

1183

August 12th, 2008 13:00

vmunix: SCSI: Read error -- dev: b 31 0x1be500, errno: 126, resid: 16384

Hi!

DMX 3
hp-ux 11.11
rp3440
brocade san switch

This error is reporting in our syslog.log and I'm thing changing the timeout is the solution.

Has any come across this before?

What is recommended timeout for LVM on HP-UX 11.11?

root@asap87[/root]: pvdisplay /dev/dsk/c27t14d5
--- Physical volumes ---
PV Name /dev/dsk/c27t14d5
PV Name /dev/dsk/c24t14d5 Alternate Link
VG Name /dev/emcvg
PV Status available
Allocatable yes
VGDA 2
Cur LV 19
PE Size (Mbytes) 8
Total PE 5393
Free PE 0
Allocated PE 5393
Stale PE 0
IO Timeout (Seconds) 90
Autoswitch On
Proactive Polling On

August 12th, 2008 14:00

Actually, I'm thinking that changing the timeout is the solution - Doh

2 Intern

 • 

1.3K Posts

August 12th, 2008 18:00

90 seconds for SYM and 180 seconds for Clarrion is enough and u r already at 90.

is the error still coming? Looks like a read error from the disk? is that a protected device?which raid level?

what path management software do you use?

6 Operator

 • 

2.8K Posts

August 13th, 2008 00:00

It looks like you don't have powerpath .. Without powerpath PVtimeout should be 180 .. Please refer to emc58837 if you need further details.

If you still see timeouts after changing timeout please file a SR. Our Customer service will be quicker in solving the issue :D

-s-

PS I still have a 3688 in my pocket .. do you swap it with a newer model ?? ;-)

6 Operator

 • 

2.8K Posts

August 13th, 2008 06:00

If you have PP timeout should stay at 90 .. But if you have PP you don't have to add the Alternate Link to your VGs ..

Open a SR .. it's easier to have a good answer quiclky IMHO.

-s-

August 13th, 2008 06:00

raid 5
powerpath

changing timeout to 180 did not solve the problem.

Any ideas?

2 Intern

 • 

1.3K Posts

August 13th, 2008 12:00

i would recommend keeping the alt link to VG even if PP is there and that is the practice i have seen. I got the logic why Stefano says the other way.. still...

6 Operator

 • 

2.8K Posts

August 25th, 2008 14:00

When you have a little spare time, can you please update us and post news on the issue ?? ThX :D

August 26th, 2008 05:00

While this problem is still a work in progress that's awaiting an OK from the user's, it appears that a bad HBA on the server is the issue. The problem is that neither HP-UX diagnostics nor Brocade nor EMC DMX 3 diagnostics, are picking up anything. Patching is not an issue on the HP-UX side. In my opinion new technology has surpassed current levels of diagnostic reporting and some new code needs to written. New code that's a lot more informative. For nothing in the HBA utility 'fcmsutil' is being reported except 'Spurious Interrupts'. And these are appearing consistently on every box.

I'm going to keep this open for a bit longer and post the results no later than soon after 9/7. 9/7 is our current window.

6 Operator

 • 

5.7K Posts

August 31st, 2008 12:00

3688.... is that some ancient Nokia thingy ?

6 Operator

 • 

2.8K Posts

August 31st, 2008 12:00

Oh no .. it's an ancient Motorola thingy :D .. Tiny and unfriendly (just like almost any other Motorola of late 90's, like StarTACs and similar phones... However they are what I call a cell-phone .. A keyboard, a tiny display, an antenna and nothing more .. The bare minimum needed to place calls (and eventually send/receive SMS) :D

6 Operator

 • 

5.7K Posts

August 31st, 2008 13:00

OMG, that's exactly what my wife wants, but what I won't buy her, since I'd like her to start emailing on her phone as well ;)
No Events found!

Top