IBM Support

Solving Big SQL High Availability Setup issues in v4.2 involving db2haicu while adding a 2nd head node - Hadoop Dev

Technical Blog Post


Abstract

Solving Big SQL High Availability Setup issues in v4.2 involving db2haicu while adding a 2nd head node - Hadoop Dev

Body

Db2haicu is used internally by the Big SQL High Availability during the install setup phase. Occasionally when trying to enable Big SQL High Availability in this phase you may see errors reported that involve db2haicu.

Below is a list of possible fixes when db2haicu is throwing an error during the add 2nd head node phase. To verify if any of these issues have worked you can manually issue the cmd after trying each possible solution
as user bigsql

db2haicu -delete;
echo $?

and verify the return code, it should always return 0 when db2haicu is working as normal. Once this command returns 0, you can rerun the Big SQL High availability add 2nd node install once again.

NOTE: This is the only time you should run db2haicu -delete, if you run it at any other time outside of the ‘add 2nd head node phase’ and you will destroy your Big SQL High availability setup.


Disable Ipv6

Ipv6 is not a recommended network configuration for Big SQL. To check if it is configured for your Ethernet adapter try the following.

ifconfig | grep -i inet6

if you find an entry for your Ethernet adapter containing ‘inet6’, then ipv6 is enabled.

To disable ipv6 you can do the following
add this line to /etc/sysctl.conf to disable all ipv6

net.ipv6.conf.all.disable_ipv6 = 1

or disable a specific Ethernet adapter ( e.g eth77 )

net.ipv6.conf.lo.disable_ipv6 = 1
net.ipv6.conf.eth77.disable_ipv6 = 1

To activate these changes run the following or simply reboot the node in question
sudo sysctl -p /etc/sysctl.conf


Verify Node host names

Db2haicu needs to be able to verify the host names of the nodes. You should verify that your node is setup to produce this information correctly by checking the consistency of the following

Check the contents of /etc/hosts on both head nodes and verify names are accurate
e.g. 9.30.104.53 mynodename.ibm.com mynodename

Check the output of the command ‘hostname‘ and ‘hostname -f‘ on both nodes and make sure they match the contents of /etc/hosts


Verify Big SQL xml files

If db2haicu is failing while trying to create TSA resources as opposed to when running with the delete option it is possible that an error occurred while creating the host xml files that bigsql generates automatically. These files are on both head nodes under

/var/lib/ambari-agent/cache/stacks/BigInsights/4.2/services/BIGSQL/package/scripts/*.xml

There will be a single xml file in this location on both head nodes after you try to enable the second head node. This file will contain the host names and other network information that bigsql has gathered from your system, verify that the information in the xml files are is correct and matches the information in /etc/hosts and the output of ‘hostname -f‘ .


Verify iTCO Modules

Db2haicu will not work if the following modules are loaded

iTCO_wdt
iTCO_vendor_support

To check for these modules do the following
lsmod | grep -i iTCO_wdt
lsmod | grep -i iTCO_vendor_support

To remove these modules do the following
modprobe -r iTCO_wdt
modprobe -r iTCO_vendor_support



If none of the above issues fix your db2haicu problems consider rebooting the node in question to reset your network adapter.

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSCRJT","label":"IBM Db2 Big SQL"},"Component":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

UID

ibm16259913