Re: [smartmontools-support] problems with CCISS as Raid 1+0 (4 disks)

[Andrei S. <as...@gm...>]

I am sure I'm using all 4 HDD sockets for Raid
.
smartctl -a -d cciss,4 /dev/cciss/c0p0
smartctl -a -d cciss,5 /dev/cciss/c0p0
smartctl -a -d cciss,6 /dev/cciss/c0p0

didn't help either.


Andrei.

---------- Forwarded message ----------
From: Andrei Sereda <as...@gm...>
Date: Dec 19, 2006 4:52 PM
Subject: Re: [smartmontools-support] problems with CCISS as Raid 1+0 (4 dis=
ks)
To: Michael Mansour <mi...@np...>


Mike,

what Smart Array Controller are you using (version, firmware etc.).
Also what are your HDDs (manufacter, size, interface etc.)?

I have used compaq arrayprobe to find out if there are  any problems (
http://www.strocamp.net/opensource/arrayprobe.php ) but seems to be
OK.  Unfortunately it doesn=B4t tell you the type and number of discs in
RAID configuration.

I can confirm this problem on 2 DB servers (ProLiant DL 585) with same
RAID configuration (4 x Seagate 15K 300G;  RAID 0+1). however it works
fine on our blades (BL 25p) with RAID 1 ( 2 x Seagate 10K 300G).
Besides the difference in RAID configuration the controllers are not
the same:

- DL586    :  Smart Array 5i Version: 2.62
- BL 25p   :  HP SA 6i Version 2.68

So it is difficult to tell what is the problem: CCISS driver,
smartmontools bug, RAID controller ?

Thanks,
Andrei.

On 12/19/06, Michael Mansour <mi...@np...> wrote:
> Hi Andrei,
>
> > Hello,
> >
> > I have HP DL585 DB server with 4 HDDs installed as Raid 1+0 on Smart
> > Array controller (Device: COMPAQ   Smart Array 5i Version: 2.62). I
> > can easily access/test cciss,0 and cciss,1 (first 2 drives in array):
> >
> > - smartctl -a -d cciss,0 /dev/cciss/c0d0
> > - smartctl -a -d cciss,1 /dev/cciss/c0d0
> >
> > When queing the other 2 I get the following error:
> >
> > # smartctl -a -d cciss,2 /dev/cciss/c0d0 (or smartctl -a -d cciss,3
> > /dev/cciss/c0d0)
> > smartctl version 5.37 [x86_64-redhat-linux-gnu] Copyright (C) 2002-6
> > Bruce Allen Home page is http://smartmontools.sourceforge.net/
> >
> > Device: COMPAQ   Smart Array 5i Version: 2.62
> > >> Terminate command early due to bad response to IEC mode page
> > A mandatory SMART command failed: exiting. To continue, add one or
> > more '-T permissive' options.
> >
> > This is the log file  (/var/log/smartd/smartd.log)
> > Dec 18 15:17:01 lnx smartd[16124]: Device: /dev/cciss/c0d0
> > [cciss_disk_02], opened
> > Dec 18 15:17:01 lnx smartd[16124]: Device: /dev/cciss/c0d0
> > [cciss_disk_02], Bad IEC (SMART) mode page, err=3D5, skip device
> >
> > I am using the latest CVS version as of today (18 Dec 06).
>
> I've exeperienced this exact same problem on Proliant BL40 blades using t=
he
> latest smartmontools CVS also. 4 disks in raid 5, first two can be seen b=
ut
> the next two can't.
>
> I haven't reported this problem yet to the list as I'm still
> troubleshooting, because the allocation of what I see from smartmontools
> doesn't make sense in my scanerio.
>
> To explain a bit, what I can see from using:
>
> smartctl -a -d cciss,0 /dev/cciss/c0d0
> smartctl -a -d cciss,1 /dev/cciss/c0d0
>
> is the output of first two disks, yet when I boot the server and enter th=
e
> SMART Array bios screen, I see:
>
> slot 0   MISSING
> slot 1   36.4gb
> slot 2   36.4gb
> slot 3   36.4gb
>
> Yes, "missing" is what I see for the first disk. If the disk is in fact
> missing, then why does it correctly display using "cciss,0" is what I fin=
d
> baffling?
>
> I get the same output as you do for cciss,2 and cciss,3.
>
> I had an engineer go to the data centre yesterday to physically look at t=
he
> server to see if this is a failed or missing disk (maybe the SMART
> controller is not reporting the disk properly).
>
> So from my setup, I experienced this on two BL40's same config.
>
> I'm now on leave from work till 2nd of January and have passed the proble=
m
> to a colleague who's handling it, so won't find out what happened till ea=
rly
> next year.
>
> What I'm saying is that I've experienced the exact same problem, but beca=
use
> my results are so far inclusive (whether it smartmontools or the smart
> controller or ?? ) I didn't think reporting this as a smartmontools probl=
em
> was the right thing to do.
>
> It looks like I'm now not the one, and if I have got a missing disk in sl=
ot
> 0, then cciss,0 should be outputting that or maybe it's querying somethin=
g
> different? (slot assignments wrong etc).
>
> Regards,
>
> Michael.
>
> > Any help will be appreciated.
> >
> > Thanks,
> > Andrei.
> >
> > -----------------------------------------------------------------------=
--
> > Take Surveys. Earn Cash. Influence the Future of IT
> > Join SourceForge.net's Techsay panel and you'll get the chance to
> > share your opinions on IT & business topics through brief surveys -
> >  and earn cash http://www.techsay.com/default.php?
> page=3Djoin.php&p=3Dsourceforge&CID=3DDEVDEV
> > _______________________________________________
> > Smartmontools-support mailing list
> > Sma...@li...
> > https://lists.sourceforge.net/lists/listinfo/smartmontools-support
> ------- End of Original Message -------
>
>