Fusion comes with tools for generating diagnostic reports that you can send to the Lucidworks Support team for troubleshooting.
Note
|
These reports are designed for interpretation by Lucidworks team members. They are not intended for self-diagnosis. |
Generating Reports in Unix
The diag.bash
script runs locally on a Fusion host and produces a compressed tarball that you can send to Lucidworks Support.
Prerequisites
The curl
utility must be installed. If it is not, you can install it like this:
-
Ubuntu
apt-get install curl
-
Redhat/CentOS
yum install curl
Additionally, the following are strongly recommended but not required:
-
Superuser access to the Fusion host.
Superuser access allows the script to collect system data. If this is not possible, execute the script as a user with sufficient permissions to access Fusion folders and the folder where the data is collected (
/tmp/diag
by default). -
An admin-level Fusion login.
The script can only collect the Fusion configuration data accessible to the Fusion user whose credentials are provided.
-
Installed on the Fusion host:
-
sysstat
This utility maximizes system data collection by the script.
-
Java JDK
The JDK, in conjunction with enabling
jps
andjstack
, is used to gather thread dumps for Fusion’s Java processes.
-
-
Enabled on the Fusion host:
-
iostat
andsar
data collection.These maximize system data collection by the script.
-
jps
andjstack
utilities in the executables path.These utilities, in conjunction with the JDK, are used to gather thread dumps for Fusion’s Java processes.
-
Running the Unix Diagnostic Script
Run the script like this, with no command-line parameters:
$FUSION_HOME/scripts/diag/diag.bash
The script prompts you to enter Fusion user credentials and a location for the output. When finished, it displays the location of the resulting tarball.
Send the tarball to Lucidworks Support for help.
Generating Reports in Windows
The diag.ps1
script runs locally on a Fusion host and produces a .zip file that you can send to Lucidworks Support.
Prerequisites
The following are required on the Fusion host where the script runs:
-
64-bit Windows Server 2008 R2 or 2012 R2.
-
curl.exe
The
curl.exe
utility depends on MS Visual C++ redistributable library. If it is not installed on your server, you can download one here. Install the redistributable, then double-click onfusion/<vers>/apps/spark/hadoop/bin/winutils.exe
from Windows explorer and make sure the OS doesn’t complain about missing DLLs. Then relaunch the shell. -
jq.exe
-
Synchronous-ZipAndUnzip.psm1
(a third-party PowerShell module)
Additionally, the following are strongly recommended but not required:
-
Administrator access to the Fusion host.
Administrator access allows the script to collect system data. If this is not possible, execute the script as a user with sufficient permissions to access Fusion folders and run system utilities like
msinfo32
andsysteminfo
. -
An admin-level Fusion login.
The script can only collect the Fusion configuration data accessible to the Fusion user whose credentials are provided.
Running the Windows Diagnostic Script
Run the script like this, with no command-line parameters:
-
cd $FUSION_HOME/scripts/diag/win64/
-
powershell -ExecutionPolicy Bypass .\diag.ps1
The script prompts you to enter Fusion user credentials and a location for the output. When finished, it displays the location of the resulting .zip file.
Note
|
Be sure to enter a fully-qualified path to the output directory. For example, enter c:\abc instead of \abc .
|
Send the .zip file to Lucidworks Support for help.