devfsadm on live system

[kouchpotato]

Hi,

I have had an a5200 disk go offline for some reason. I would like to perform a devfsadm -Cv to clean up the hanging logical links.
This is a live mission critical box and I'm in the nasty stuff if it goes wrong.
Can I perform a devfsadm -Cv on a live box? Whats the consensus?

thanks

K

 

[daFranze]

welcome to the club! I recently fixed a photon offline problem.

I do not recommend to run devfsadm in a live environment; I guess It will work, but you will never know.

What exactly is the problem? why do you want to clean logical links?

A5200 Problem: what did you change? what is the problem?

Regards
-- Franz
Sorry I'm not a native spaeker, I'm from Munich, Germany - "Home of the Whopper", oh no, "Home of the Oktoberfest" ;-)
Solaris System Manager; I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[kouchpotato]

Franz,

If I run a format on a v880 system, the command will hang for a number of minutes and returns all internal disks and only 2 of the 22 disks in the a5200.
The a5200 disks report <drive not available: formatting>.

I suspect that one of the disks on the a5200 has reached capacity and offlined. The a5200 itself shows no errors.

The a5200 is setup as single loop.

I am new to FC and luxadm so be gentle!

K

 

[daFranze]

do you see any errors/infos in /var/adm/messages?
can you see the photon in 'luxadm probe'? can you see disks in 'luxadm display boxname'
Where do you have disks in the photon (which slot)? There are some minimum requirements...

Regards
-- Franz
Sorry I'm not a native spaeker, I'm from Munich, Germany - &quot;Home of the Whopper&quot;, oh no, &quot;Home of the Oktoberfest&quot; ;-)
Solaris System Manager; I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[kouchpotato]

in messages file is [ID 686697 kern.info] NOTICE: Qlogic qlc(1): Loop OFFLINE - This is approx the time database crashed as well.

There follows a number of offlining events ..

'luxadm probe' finds 2 devices

'luxadm display FCloop' displays 11 devices.

The Photon has all 22 slots allocated.

K

 

[daFranze]

did you setup this Box recently or was it in use for a long time?
Loop online/offline problems are wellknown to Sun; I recommend to open a Call at Sun, they will check the Revisions of Disks, A5x00, FC-HBA, Hostdrivers and recommend to install a set of patches; If this does not fix they will send a Technican with "Store Tools" Software and this guy will start with a minimum configuration and will plugin one device after another. This may take some time...

I think deleting symbolic links will not fix the Problem with slow format etc. Whenever the OS sees the Disks devfsadmd will create a link faster than you can delete them...

Regards
-- Franz
Sorry I'm not a native spaeker, I'm from Munich, Germany - &quot;Home of the Whopper&quot;, oh no, &quot;Home of the Oktoberfest&quot; ;-)
Solaris System Manager; I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[daFranze]

btw: I am using Sun's Storage Automated Diagnostic Environment 2.3; this Tool will watch all FC Loops, Disks, GBICs, IBs etc; and sends a mail if there is something wrong...

Regards
-- Franz
Sorry I'm not a native spaeker, I'm from Munich, Germany - &quot;Home of the Whopper&quot;, oh no, &quot;Home of the Oktoberfest&quot; ;-)
Solaris System Manager; I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[kouchpotato]

Thanks for your Help Franz ..

The photon is a new install .. was up and running for about a week before the problem.

Have been trying to find patch info for the a5200 .. I think this may well be where the problem lies.

I have a Sun Techy on site at the moment - will see what he has to say on the matter.

once again thanks for your help

K

 

[daFranze]

what did they find out? I guess they changed GBICs and installed a bunch of patches, right?! ;-)

Regards
-- Franz
Sorry I'm not a native spaeker, I'm from Munich, Germany - &quot;Home of the Whopper&quot;, oh no, &quot;Home of the Oktoberfest&quot; ;-)
Solaris System Manager; I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[Scunningham99]

Problem was due to faulty HBA card... This is now resolved...


Sy UK