Cisco ISE Upgrade URT error

[sameerasripad]

Dear Team,

We are having distributed deployment of ISE nodes for RADIUS that is 2 PSN, 2 PAN and 2 Monitoring nodes. We are planning to upgrade the firmware from 2.6.0.156 Patch 10 to 3.1.0.518 Patch 4.

I was running URT application on Secondary administrative node and I have come across error [opt/urt/estimate_time.sh: line 18: ERROR:ORA-12545: Connect failed because target host or object does not exist]. I can see in some of the articles mentioned as bugs.

Does this error really impact the upgrade. I am planning to upgrade the nodes using GUI mode.

Please find the attached logs for reference.

[Thomas Schmitt]

Hi,

you can take a look in your estimated_time.sh file.You can unzip urtbundle multiple times with some archive program, until you can see scripts.

setValues()
{
TOTAL_TIME=0
PRETIME=25
BKPTIME=$EXPORTTIME
RPMSTIME=30
OSTIME=60
FROMVERNODOTS=`echo $FROMVERSION | tr -d '.'`
if [ `cat /tmp/node_info.txt | grep ":STANDALONE:" | wc -l` -eq 1 ] ; then
  MNTSIZE=`echo -e " set pagesize 0; \nset head off ; \n  set trimspool on feedback off; \n SELECT ltrim((round((sum(bytes)/1024/1024/1024),0)))  from dba_segments where owner='MNT'; " | su - oracle -c "sqlplus -S /@cpm10"`
else
	MNTNODES=`cat /tmp/node_info.txt | grep MNT | awk -F":" ' { print $1 } '`
    MNTNODE=$(awk -F" " '{print $1}' <<<$MNTNODES)
MNTSIZE=`echo -e " set pagesize 0; \nset head off ; \n  set trimspool on feedback off; \n SELECT ltrim((round((sum(bytes)/1024/1024/1024),0)))  from dba_segments where owner='MNT'; " | su - oracle -c "sqlplus -S cepm/$PASSWD@'(DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = $MNTNODE)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SID = cpm10) ) )'"`
fi
MNTTIME=$(($MNTSIZE*5))
MNTTIME=$(($MNTTIME+5))
if [ `echo $MNTTIME | grep -c ORA-` -ne 0 ]
then
        logecho_urt "  - Failed in getting mnt size details from DB"
        exit 1
fi
}

So probably HOST here isn't set

MNTSIZE=`echo -e " set pagesize 0; \nset head off ; \n  set trimspool on feedback off; \n SELECT ltrim((round((sum(bytes)/1024/1024/1024),0)))  from dba_segments where owner='MNT'; " | su - oracle -c "sqlplus -S cepm/$PASSWD@'(DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = $MNTNODE)(PORT = 1521)) (CONNECT_DATA = (SERVER = DEDICATED) (SID = cpm10) ) )'"`

MNTNODE set fails probably somewhere here

MNTNODES=`cat /tmp/node_info.txt | grep MNT | awk -F":" ' { print $1 } '`
    MNTNODE=$(awk -F" " '{print $1}' <<<$MNTNODES)

Just check your error in your environment.

[sameerasripad]

Dear Thomas,

Thank you for your reply.

Could you please let me know the path to find estimated_time.sh file as I don't see any error logs being saved in the directory. 

[Thomas Schmitt]

its part of urt bundle

[thomas]

Sounds like a great question for the TAC.

[Raghas Naveed]

I am facing the same issue, can you please share your experience, how it went for you?

[Damien Miller · Accepted solution]

Issues encountered while running the URT bundle should really be looked at by TAC with case. There are way too many variables for a forum user to give accurate information as to why it might be failing. They will need to pull logs. 

[Raghas Naveed]

I was facing the same issue while running URT bundle on SAN. The time estimation for MnT node failed because they had huge log files (1+ TB). I did the upgrade with CLI method. The URT error was a cosmetic one, got it checked via TAC. The upgrade went smoothly without issues, but took around 14+ hours for only one MnT node and we did it for 2 Admin Nodes , 2 MnT nodes and 28 PSNs. 

Admin Ones took 2 hours each, MnT took 14+ Hours each and Each PSN took around 1-2 hours to upgrade.