Changing Key and Trust Store Passwords
Change key and trust store passwords by using the
${MAPR_HOME}/server/manageSSLKeys utility.
Release 7.0.0 added a new changepassword command to the
${MAPR_HOME}/server/manageSSLKeys utility. The existing
copywithconfiguredpassword and createrandompassword
commands remain for upgrade purposes but are deprecated starting with release 7.0.0.
-k option. To change the trust store password, you must provide the current
trust store password with the -t option. To set the new user-selectable
password, use the -kp or -tp option. Otherwise, a random
password is created. Note that you must pair the -kp and/or
-tp options with the -k and/or -t
options, respectively. For
example:# /opt/mapr/server/manageSSLKeys.sh changepassword \
-k 8zVMhs8RtLDXpnTTIBqQkt_q_pFFV3I_ \
-t 5eqHoTrLRaiev6dwxJhfzm3qpPqW_0J2 - Run the
manageSSLKeys.sh changepasswordcommand on the first node in the cluster. Running the command creates a directory under/tmp, with new password files and a script. A newstore-passwords.txtis also created in this directory. It is a best practice to keep the passwords in this file and deletestore-passwords.txtfrom the/tmpdirectory. - Stop ZooKeeper and Warden on all nodes in the cluster.
- Distribute the above directory to all nodes in the cluster.NOTEInstead of distributing the directory to all nodes in the cluster, run the
manageSSLKeys.sh changepasswordcommand used in step 1 on each node. This option eliminates file type and format issues in a cluster on both FIPS and non-FIPS nodes. - On each node in the cluster, make sure they have the correct ownership and permissions,
and then run
copyPasswordFiles.shfrom this directory. - Run
configure.sh -Ron all nodes to allow all services to update their configuration. - Start ZooKeeper and Warden on all nodes in the cluster.
The security-file type and format are different on FIPS- and non-FIPS-enabled nodes. You cannot copy the modified passwords across FIPS to non-FIPS or vice versa. To change a password with both FIPS and non-FIPS nodes in a cluster, run the procedure twice: once on the FIPS node and once on the non-FIPS node. Only copy the generated files to, and run the script on, nodes with the same FIPS or non-FIPS type.