Hi Anton,

how about:

rm /dev/vx/dmp/*
rm /dev/vx/rdmp/*

Regards
Przemek

On Fri, Oct 30, 2009 at 06:06:07PM +0300, Antony Pavlenko wrote:
> Hello.
> This procedure is online.
> I've tested it now more then 10 times "on production server where lots of 
> Oracle databases are working".
> 
> Anton Pavlenko
> 
> 
> On Fri, Oct 30, 2009 at 01:29:58PM +0100, przemol...@poczta.fm wrote:
> > Hi all,
> > 
> > it happened on production server where lots of Oracle databases are working 
> > all the time
> > and finding maintenence window longer then 15 minutes is ... :-(
> > let alone rebooting server.
> > 
> > Any _online_ procedure ?
> > 
> > Regards
> > Przemek
> > 
> > On Fri, Oct 30, 2009 at 01:15:22PM +0300, Antony Pavlenko wrote:
> > > It is a new feature :-)
> > > You need to cleanup solaris device tree  after disk removal and before 
> > > you add any new disks if you want to use the same host LUN. 
> > > 
> > > After some experiments and reboot as the results of this experiments we 
> > > have found correct cleanup procedure.
> > > 
> > > For me it works and after this new devices can be added with aot of
> > > reboot:
> > > 
> > > *remove disks from all disk groups
> > > * then remove disks from veritas ( vxdisk rm <disks> )
> > > * offline all paths to this disks
> > > luxadm -e offline /dev/rdsk/......
> > > 
> > > I'm doing it like this :
> > >  vxdisk path | awk ' $2 ~ /DISKS_REGEXP/  { print "/dev/rdsk/"$1 } ' > 
> > > /tmp/all_disks_paths
> > > # vxdisk list | awk ' $0 ~ /DISKS_REGEXP/  { print $1 } ' | while read # 
> > > disk; do vxdisk rm $disk; done
> > > cat /tmp/all_disks_paths | while read disk; do luxadm -e offline $disk; 
> > > done
> > > 
> > > Than you need to unconfigure this disconnected devices with cfgadm ( 
> > > cfgadm -c unconfigure -o unusable_SCSI_LUN or unusable_FCP_dev, depends 
> > > of solaris version ).
> > > 
> > > All paths can be found like this :
> > > cat /tmp/all_disks_paths  | nawk ' { str=substr($1,11); 
> > > sub(/c[0-9]*/,"&::",str);  sub(/d[0-9]*s2/,"",str); sub("t","",str); 
> > > print tolower(str); } ' | sort | uniq | while read path; do  cfgadm -c
> > > unconfigure -o unusable_SCSI_LUN  $path; done
> > > 
> > > And then
> > > 
> > > devfsadm -Cv
> > > 
> > > After this cleanup veritas tree :
> > > 
> > > vxddladm stop eventsource
> > > mv /etc/vx/disk.info /etc/vx/disk.indo_backup
> > > mv /etc/vx/array.info /etc/vx/array.info_backup
> > > rm /dev/vx/dmp/*
> > > rm /dev/vx/rdmp/*
> > > vxdisk scandisks
> > > vxdctl enable
> > > vxddladm start eventsource
> > > 
> > > And now You can add new devices :-)
> > > 
> > > 
> > > With regards,
> > > Anton Pavlenko
> > > 
> > > On Tue, Oct 27, 2009 at 08:00:29PM -0400, Ram Rudra wrote:
> > > >    I had the same issue on a CVM/CFS cluster node and the issue was not
> > > >    resolved with all the recommended steps.
> > > >    Apparently the error disappeared after the reconfiguration reboot of 
> > > > the
> > > >    node. Not sure if this is a new bug in 5MP3RP1?
> > > > 
> > > >    Rudra
> > > > 
> > > >    2009/10/27 <[1]przemol...@poczta.fm>
> > > > 
> > > >      On Mon, Oct 26, 2009 at 02:41:40PM +0100, [2]przemol...@poczta.fm 
> > > > wrote:
> > > >      > Hello,
> > > >      >
> > > >      > I am trying to add to a server new LUNs but face the following
> > > >      problem:
> > > >      >
> > > >      > server:/root# vxdctl enable
> > > >      > VxVM vxdctl ERROR V-5-1-0 ?Data Corruption Protection Activated 
> > > > - User
> > > >      Corrective Action Needed
> > > >      > VxVM vxdctl INFO V-5-1-0 ?To recover, first ensure that the OS 
> > > > device
> > > >      tree is up to date (requires OS specific commands).
> > > >      > VxVM vxdctl INFO V-5-1-0 ?Then, execute 'vxdisk rm' on the 
> > > > following
> > > >      devices before reinitiating device discovery:
> > > >      > ? ? ? ? c5t0d25, c5t0d25
> > > >      > server:/root# vxdisk rm c5t0d25
> > > >      > server:/root# vxdctl enable
> > > >      > VxVM vxdctl ERROR V-5-1-0 ?Data Corruption Protection Activated 
> > > > - User
> > > >      Corrective Action Needed
> > > >      > VxVM vxdctl INFO V-5-1-0 ?To recover, first ensure that the OS 
> > > > device
> > > >      tree is up to date (requires OS specific commands).
> > > >      > VxVM vxdctl INFO V-5-1-0 ?Then, execute 'vxdisk rm' on the 
> > > > following
> > > >      devices before reinitiating device discovery:
> > > >      > ? ? ? ? c5t0d25, c5t0d25
> > > >      > server:/root# vxdisk rm c5t0d25
> > > >      > ...(and again the same)...
> > > >      >
> > > >      > Can anybody help me ?
> > > > 
> > > >      I suspect it happened just after we have moved (online) array LUN 
> > > > to a
> > > >      new place using arrays feature.
> > > > 
> > > >      Regards
> > > >      Przemyslaw Bak (przemol)
> > > >      --
> > 
> > -- 
> > Regards
> > Przemyslaw Bak (przemol)
> > --
> > http://przemol.blogspot.com/
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 

-- 
Regards
Przemyslaw Bak (przemol)
--
http://przemol.blogspot.com/





























Tanie rozmowy telefoniczne!
Sprawdz >> http://link.interia.pl/f2410

_______________________________________________
Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx

Reply via email to