09-30-2011 05:56 AM
Hello,
While running the vxfentsthdw, I see this call chk_dd_process() failing. I wish to know what this call does and what it is expecting.
Thank you,
Sumana
10-02-2011 08:41 AM
Sumana,
Please let me know what version of Veritas you are using.
If some disk fails for vxfentsthdw test, reason could be fencing not configured properly or disk is not SCSI-3 compatible.
Below post may help you.
https://www-secure.symantec.com/connect/forums/test-disk-io-fencing-vxfentsthdw#comment-4092181
Regards,
~Anoop
10-05-2011 09:52 AM
Hello Anoop,
Thank you for your reply.
This is what I am using.
# $Id: vxfentsthdw_linux.sh,v 1.31 2009/06/25 12:16:38 ssharma3 Exp $ #
I have implemented the reservation commands on the Disk array according to the specs. For the preempt and the preempt_abort, I did have a win2k validation go through without errors. However I am seeing problems with the vxfentsthdw tool. The chk_dd_process seems to pass, but there is an error message from this line 'VXFENLOG "chk_dd_process() failed for $1\n"'
I have taken care to see that the IOs on the preempted nexus has been aborted and the registrations/reservations deleted.
Please help me on this. I am new to the IO fencing concept and so not sure what I need to check.
chk_dd_process()
{
pre_dot_text="Test to see if I/O on node $1 terminated"
pad_dots $pre_dot_text
VXGETTEXT "Test to see if I/O on node $1 terminated $dots \c" 3032 "$1" "$dots"
count=0
while [ $count -lt 300 ]
do
sleep 1
if localhost $1
then
$PS -ef| $GREP -v grep | $GREP "$DD" 1>/dev/null 2>&1
echo $? > $output 2> /dev/null
else
$RSH $RSH_OPT1 $1 $RSH_OPT2 "$PS -ef" | $GREP -v grep | $GREP "$DD" 1>/dev/null 2>&1; echo $? > $output 2> /dev/null
fi
status=`$CAT $output`
if [ $status -eq 0 ]
then
count=`expr $count + 1`
#echo "count=" $count
continue
else
VXGETTEXT "Passed" 3002
VXFENLOG "chk_dd_process() failed for $1\n"
break
fi
done
#echo "count=" $count
if [ $count -eq 300 ]
then
VXGETTEXT "Failed" 3003
VXLOGTEXT "vxfentsthdw: Test failed from node $1"
VXGETTEXT "\nI/O not killed after preempt abort of key KeyA on node $node1\n\tDisk is not SCSI-3 compliant on preempt_abort command.\n\tExecute the utility vxfentsthdw again and if failure persists contact\n\tthe vendor for support in enabling SCSI-3 persistent reservations\n" 3033 "$node1"
VXFENLOG "chk_dd_process() passed for $1\n"
cleanup_exit
exit 1
fi
}
Thank you!
Sumana
10-05-2011 10:01 AM
Hello Anoop again,
Forgot to mention something strange I noticed. In the script vxfentsthw, in every call the passed case looks like different from the chk_dd_process call. Did the call pass or fail? How are there 2 log messages in the same case with conflicting terms.
thank you again
Sumana
VXGETTEXT "Passed" 3002
VXFENLOG "preempt() passed for $1\n"
VXGETTEXT "Passed" 3002
VXFENLOG "delete_keys() passed for $1\n"
VXGETTEXT "Passed" 3002
VXFENLOG "clear() passed for $1\n"
VXGETTEXT "Passed" 3002
VXFENLOG "preempt_abort() passed for $1\n"
VXGETTEXT "Passed" 3002
VXFENLOG "chk_dd_process() failed for $1\n"
10-06-2011 11:15 AM
VRTSvcs-5.1.132.000-SP1RP2_RHEL5
Cheers
Luis
10-07-2011 03:56 AM
Sumana,
I am not sure on sure on code level and may need more time to understand and advise.
Have you opened a case with Symantec ? It could be related to some bug and fix available with them.
10-08-2011 02:38 AM
Thank you Anoop,
I have contacted the customer service, will post the feedback once I hear from them.
Cheers,
Sumana
10-10-2011 02:31 AM
Hello Anoop,
The customer care has redirected me to the same forum in turn. Could you please direct me to the right channel to get this sorted?
Thank you,
Sumana
10-10-2011 07:20 AM
I am wondering why they directed you to forum. Do you hav proper support enable for this product ?