Upgrading Analyzer probe server
If you upgrade the Analyzer probe server, the RAID Agent on the same host is also automatically upgraded.
Prerequisites for upgrading Analyzer probe server
Before you upgrade the Analyzer probe server, ensure that the system-wide and user-level limits on the number of open files is set on the Linux host.
- Minimum values for system-wide and user-level: 65535
- Recommended values:
- System-level: 327675
- User-level: 262140
For more information, refer Increasing the maximum number of open files.
Upgrading the Analyzer probe server
The installer (dcaprobe_install.sh
) starts
and stops the crond service. Therefore, do not run any operations that use the crond service
when the installer is running.
Before you begin
- To upgrade the Analytics probe server from a version earlier than 4.0.0, you must first upgrade the Analyzer probe server to version 4.0.0.
- A license for the Analyzer probe server must be registered.
- Review the Analyzer probe server requirements (hardware and software).
- When upgrading from a version earlier than 10.0.0, make sure that both the root directory and the installation directory of the host on which you plan to install the Analyzer probe server has 5 GB of free space.
- During the upgrade, /opt/jp1pc/htnm/HBasePSB/hjdk/jdk might be deleted. If you have created files under this directory, move them elsewhere. If any settings (such as htnm_httpsd.conf) reference a file under this directory, revise them to use the new location.
- Log on to the host where the component to upgrade is installed.
- Stop any security monitoring software, antivirus software, and process monitoring software.
- Mount the Hitachi Ops Center installation media and copy the directories and files in the DCAPROBE directory from the installation media to a directory on the Linux host. Note
- You must use only the following characters in the directory path to which the
installer is copied:
A
-Z
a
-z
0
-9
-
.
_
- Do not use spaces.
In the following example, if the
/root/DCAPROBE
directory already exists, create a new directory, and then perform the subsequent steps in the new directory.mkdir /media/OpsImage mount /dev/cdrom /media/OpsImage mkdir /root/DCAPROBE cp -rT /media/OpsImage/DCAPROBE /root/DCAPROBE
- You must use only the following characters in the directory path to which the
installer is copied:
- Move to the
/root/DCAPROBE
directory.cd /root/DCAPROBE
- Run the precheck tool as a root user to check whether Analyzer probe server can be installed:
sh ./dcaprobe_precheck.sh
If
OK
is displayed in[
Check results
]
, you can start the installation. IfNG
is displayed, make sure the system requirements have been met.Output example when the Ops Center Analyzer version is 10.0.0:
============================================================ Ops Center Analyzer probe Precheck ver. 10.0.0-00 ============================================================ [ Check results ] Ops Center Analyzer probe server [10.0.0-00] [OK] [ Details ] Check resolved hostname. [host-name (IP-address)] [OK] Check premise OS version. [OK]
If the following message is displayed, you must change the JDK used by the Analyzer probe server. For details, see Resolving a JDK-related error for the Analyzer probe server.
JDK environment is invalid (invalid-settings).
For invalid-settings, one or more of the following values is displayed:java
,keytool
,jstack
,jre_1.8.0
, orjava_home
.Note- When you run the precheck tool, it checks the static information of the system environment.
- If the
-v
option is specified, information such as the installed version of Analyzer probe server and the OS name is also displayed.
- Run the following command as root to
start the upgrade:
sh ./dcaprobe_install.sh VUP
- Do not change the size of the device window while the command is running. If you change the size of the window, the installation fails.
- If the following message is displayed, you must change the JDK used by the
Analyzer probe server. For details, see Resolving a JDK-related error for the Analyzer probe server.
[ERR] JDK environment is invalid (invalid-settings).
For invalid-settings, one or more of the following values is displayed:
java
,keytool
,jstack
,jre_1.8.0
, orjava_home
.
- Refresh the browser cache.
Troubleshooting upgrade installation issues
The
megha-jetty.sh
command returns
Not running
and
Analyzer probe server upgrade installation fails.
The upgrade installation might fail due to the following issues:
Condition 1: Executing the
megha-jetty.sh command returns the message
Not running
when the
Analyzer probe server is running.
/usr/local/megha/bin/megha-jetty.sh
Condition 2:: The PID (process ID) reported by the following commands are different:
ps -ef | head -n 1; ps -ef | grep start.jar | grep /usr/local/megha
cat /usr/local/megha/misc/jetty.pid
For example:
# ps -ef | head -n 1; ps -ef | grep start.jar | grep /usr/local/megha UID PID PPID C STIME TTY TIME CMD megha 2271 1 15 13:37 ? 01:15:38 java -XX:+UseParNewGC -XX:MinHeapFreeRatio=20 -XX:MaxHeapFreeRatio=20 -Xmx15534m -jar start.jar -Dapp.home=/usr/local/megha # cat /usr/local/megha/misc/jetty.pid 31534
When both these conditions occur, run the following procedures for Analyzer probe server.
Procedure
Using an SSH client, log in to the Analyzer probe server as a root user.
Run the following command, and insert a hash mark (#) at the beginning of each line to comment out the entry:
crontab -e
#*/5 * * * * F=/usr/local/megha/cron.5min; test -f $F && bash $F #*/5 * * * * F=/usr/local/megha/bin/sysstat.sh; test -f $F && (bash $F >> /usr/local/megha/logs/sys/`date +\%Y\%m\%d`.log)
Run the following command to obtain the PID (process ID) of the jetty service:
ps -f | head -n 1; ps -f -g megha | grep start.jar | grep /usr/local/megha
Run the following command to stop the jetty process:
kill <PID>
Run the following command to confirm the jetty process is stopped.
ps -f | head -n 1; ps -f -g megha | grep start.jar | grep /usr/local/megha
Run the following command to confirm that there is no Analyzer probe process running:
ps -f | head -n 1; ps -f -g megha | grep megha
If the process is still running, run the following command using the reported PID:
kill <PID>
Run the following command to confirm the Analyzer probe process is stopped.
ps -f | head -n 1; ps -f -g megha | grep megha
Run the following commands to start the Analyzer probe server service and confirm the process is started:
/usr/local/megha/bin/megha-jetty.sh start
/usr/local/megha/bin/megha-jetty.sh status
Run the following commands and confirm that the two reported PIDs match:
ps -f | head -n 1; ps -ef | grep start.jar | grep
/usr/local/megha cat /usr/local/megha/misc/jetty.pid
For example:
# ps -f | head -n 1; ps -ef | grep start.jar | grep /usr/local/megha UID PID PPID C STIME TTY TIME CMD megha 2271 1 15 13:37 ? 01:15:38 java - XX:+UseParNewGC XX:MinHeapFreeRatio=20 - XX:MaxHeapFreeRatio=20 -xmx15534m -jar start.jar -Dapp.home=/usr/local/megha # cat/usr/local/megha/misc/jetty.pid 2271
Run the following command again and delete the hash marks (#) from the beginning of each line that you added earlier:
crontab -e
*/5 * * * * F=/usr/local/megha/cron.5min; test -f $F && bash $F */5 * * * * F=/usr/local/megha/bin/sysstat.sh; test -f $F && (bash $F >> /usr/local/megha/logs/sys/`date +\%Y\%m\%d`.log)