The BVQ Offline Scanner is the application of choice to collect information for a BVQ analysis without installing any program. Start this application on any server or laptop and let it collect configuration and performance information for some time. These information needs to be sent to a central BVQ service point (SVA or partners) where it will be analyzed for you.
The offline scanner is collecting technical data like performance data and topology information as for instance volumes, sizes of volumes, names of the volumes. Information stored on the volumes is not visible to the offline scanner and will therefore not be collected.
This scanner works for all SVC/Storwize versions from 6.4.10
Download the scanner and unzip it to a folder on a Microsoft Windows system with access to the SVC or Storwize system.
Please avoid to use blanks or special characters in the path to the folder. We recognized this as bug in the scanner. The scanner will not start completely in this case. This is solved and will be available soon.
Check that your system has SSH (port 22) access to the cluster IP address of the SVC or Storwize System to be scanned
Check for enough free space in the filesystem where the scanner is installed (10GiB should be sufficient).
On the scanned SVC or Storwize systems:
SVC or Storwize delivers its statistics in a configurable interval. Please adjust this accordingly:
5min intervals are best for risk or storage planning analyses
1min intervals are best for short term performance issue analysis
If the scanned SVC or Storwize systems are also scanned by IBM TPC or VSC, please change the interval in the appropriate configuration panes of these products. Changes on the SVC or Storwize systems will be reset by the TPC or VSC scanners according to their current configuration.
Provide login credentials for an administrative user. Administrative rights are only needed to perform the cpdumps command, to copy all statistics files on the config node.
Ensure that the system has an attachment to a running NTP server configured.
Step 3: Start the scanner
Some thoughts before you start the scanner
Running the scanner is easy and uncritical. No installation is needed and it can be run on any client system with TCP/IP connection to the systems to scan..
Scan time is dependent on the objective of the scan
10 -14 days of scanning is sufficient to prepare for any kind of analysis like risk analysis or IO density analysis for storage planning. This timeframe is also used for a BVQ PoC purposes.
5-7 days of scanning is the minimum to prepare for a risk analysis
1-2 days of scanning can be used for short term performance issue analysis (e.g. peak latency analysis) if a bottleneck actually happens in this time (Please tell us exactly, at which point in time your performance issue occurred)
You need to go through a sequence of dialogs to get it scanning:
Create a connection profile to be able to reuse your configuration
Enter the IP Address and the username/password (or username/SSHkey) of the SVC or Storwize system.
Survive all the prerequisite checks in the next dialog. Correct all warnings or errors showing up there.
Observe the scan process and your free space in the filesystem.
Get the resulting zip archive file.
The scanner will run until you force it to stop. During that, the scanner will continuously gather the performance statistics and configuration information.
After the scan has finished you need to upload the resulting data zip file to the BVQ service point to be analyzed there.
Step 3.1: Start the scanner
Start the scanner by just double clicking the file named bvq-svc-scanner-offline.exe.
The scanner needs an installed java runtime ("jre") to start.
Step 3.1: Define a profile per SVC or Storwize system
Offline Scanner profiles allow to run more than one scanner in parallel. This is helpful in case multiple SVC / Storwize systems are included in coupled environment like two clusters in a Metro/Global mirror partnership or Storwize systems below an SVC.
If you want to scan such a multiple SVC / Storwize System environment:
Start one scanner per system. You get a separate control window for each running instance.
Configure one profile (IP connection information and user credentials) for each system as described in "Step 1" below.
Each profile configuration is kept and can be reused.
Each profile uses its own working directory in the following path below the installation directory ./instances/<profilename>/
After finishing one instance the scanner offers to store the result ZIP archive file in the working directory with a filename including the cluster name and its alias ID.
Step 3.2 Configure system access
If you start the scanner a second time, all data data which has been entered here can be reused.
You can enter user authentication credentials either with user name and password or SSH private key file.
Please enter the max age of the data! If you intend to collect data more data than only 7 days you have to adapt this value
If you scan a V7000 unified then please take care that you enter the IP address of the storage and not the unified heads.
The offline scanner can not distinguish between these systems and will not deliver data if you connect to the unified heads.
Step 3.3: Start the automated prerequisite check
The scanner tool checks whether it can safely access the SVC / Storwize system before it starts to scan it.
The scanner is running and gathers configuration information and performance statistics and stores these in the instance working directory.
Click finish to end the scan process and to proceed to "Step 3.5" below.
Please frequently check whether the scanner is still working and is able to access the SVC / Storwize system. The safest way to do this, is to open the sub directory ./instances/<profilename>/iostats/new and check whether you find a folder which was created less than 10 minutes ago. (klick below to enlarge picture)
Step 3.5: Stop scan to store the scan result package
Finally the tool packs all gathered information in a resulting ZIP archive file. The tool defaults to store this to the instance working directory.
For example, if you named the profile "production" then you will find this ZIP archive at .\instances\production\
You should send all zip files from all scans.
Step 4: Contact us to upload the data for analysis or reporting
Please use the following request and enter 'offline scan' into the message field.
Step 5: Start a performance bottleneck analysis
The output of the BVQ Offline Scanner can now be used for any kind of performance analysis
Performance bottleneck analysis on IBM SVC and IBM Storwize V7000— Customer example: performance bottleneck due to MDG cache overflow. The following document shows how to find the root of the cause and with it the volume, which caused the overflow. Also included is SAN (node port) analysis and node plus global cache analysis.