Starting with version 3.0 the upgrade from DB2 to version 10.5 FP5 is mandatory.
Please download first this DB2 version from the IBM website and then perform the upgrade. For instructions, please see the following document: DB2 upgrade from version 9.7.4 or higher to 10.5.
- Loading and drawing processes of the performance view were significantly improved
- Long lasting data load processes can now be interrupted.
- The performance view for node ports classifies the data and IO rate now according to 'node to local node' and 'node to remote node'.
- All object types in the topology path can now be set behind the object types 'node' and 'IO group'.
- The VDisks which are potentially using a specific node port can now easily by represented by the node´s context menu in the performance view.
- SVC code version 7.5 beta toleration support
Interruption of a loading process
Long-running loads processes of the performance view, can be aborted via the red button right to the loading bar in the status line.
Classification of some node port metrics in 'node to local node' and 'node to remote node'
The performance view for node ports classifies now the data-- and command rate in 'node to local node' and 'node to remote node'. The previous existing column 'node' was replaced accordingly.
The remote data- and command rate is particularly useful in environments which are using remote copy (metro / global mirror) over long distances in order to derive directly the load on the transmission lines.
Other object types behind nodes and IO groups
In the topology path all SVC object types can now be set behind the object types 'node' and 'IO group' .
This can be useful to determine if a 'node' is reaching its performance limits in certain configuration cases.
Load of VDisks behind node ports
The load of a VDisk group runs via certain node ports into the system. Due to the different multipathing procedures, it is impossible to exactly determine which path is used by a certain VDisk.
Therefore, the amount of VDisks which are using a particular node port can only be roughly estimated . The localisation is done by the intersection of:
- Connection via access group IO ports: VDisks which are associated with an access IO group containing this node port.
- Connection via SAN-zoned host ports to node ports: VDisks whose mapped hosts have a host port in an SAN zone with this node port.
When for example high data rates or a buffer credit wait at a certain date are set for a node port and the causing VDisks have to be found, this can now be easily represented via the object context menu in an additional performance view:
|BVQ-8870||Common||SVC version 7.5 toleration support|
Performance view: node port: differentiate metric data rate and command rate node according to 'node to local node' und 'node to remote node'
Topology view: enable more paths behind 'node' and 'IO group'.
Performance view: disable the blocking of the GUI during loading process for PerfView
Performance view: enable interruption of the loading process
Performance view: acceleration of the drawing process for many simultaneous curves by revising the drawing routine
Performance view: acceleration of the drawing processs by using BVQSamplingXYLineRenderer without dots (default)
Performance view: node port: options-dialog: sorting SAS events analogously to FC events
Performance view: button "*": adjust tooltip.
Object node: detail view: show model number additionally in plain text
Topology view: search field: expand results table with 'value' column
Topologiy view: topology navigator: text mode: recognize '/' as final
Topology view: node port: enable a more consistent coloring of the states for IP (iSCSI, iPrep), PCIe and FCoE ports
Object context menu: node port: enable access to relevant VDisk performance view
DB/cluster selection dialog: allow operation even during a waiting period
Tab bar: shorten tab identifiers of the performance view
Tab bar: active tab now always visible by auto-scrolling
Debugging: show VM options in the exception handling dialog
FC and SAS ports can have the same WWPN in a cluster => persist fails
Setup installer: indicates a failure in rare cases as success
Topology persist: exception after updating to code version 18.104.22.168
Toplology persist: exception
Ignores various attributes of the VDisk copy from code version 7.3 or later.
Separator ':' causes problems when scanning Storwize systems starting with code version 7.4
Saves PI entries partially without any values
CLI doesn`t start
Schema migration: installer: update of existing installation no longer available because of missing softlink '/ logs'
Schema migration: installer: update: topology & IOstats path is not read out
Error in argument 'force update'
When creating a disc types an optional name / comment cannot be placed
SVC scanner: topology folder & IOstat folder are not adopted correctly during update
Update: for 'leave unchanged' the respective installer exe should be retained
Treemap coloring: aggregation of IO density coloration to higher lying objects (grey color)
Treemap coloring: 'Response time compliance': object type storage class: not after peak MDisk group but after peak MDisk
Treemap coloring: node port: SAS: state 'online' doesn´t change to green
Treemap: invert: labeling via SHIFT button is not working properly
Treemap: SeC sorting in the treemap is not always working properly
Treemap: aspect topology: shift-selection doesn't follow the PF (SeC)
Performance view: some keyboard shortcuts don't work properly
Performance view: options: export of aggregated charts delivers an empty csv file
Performance view: memory overflow if date is changed in large periods
Performance view: time-favorite: 'Put timing to all open' doesn't update all open performance views
Node models V7000-100 and V7000-300 are not differentiated correctly
PI loading procedure: particularly wrong results after loading past PI
DB connection profile: file doesn't remain in the previous sub-directory after edit
DB reload is not finished - GUI is locked (hourglass blocks any entry)
Topo reload doesn't execute the PI reload anymore
Teilweise wird für Objekte kein Realtime PI Timestamp angezeigt
Docking: moving of the performance view option dialog from a docking main window brings the main window to the front
Docking: now and then NPE when closing a performance view
Docking: IllegalArgumentException when closing a tab in the lower half of the main window
Docking: problems (exception) when a dock is minimized
Docking: NPE DragOperation.getCurrentMouse
Performance view: CPU core: primary column 'Core' doesn't have int-sort but alpha-sort
Performance view: 'Destage count in flight' on unit IO/s instead of 'Count'
Performance view: some lines are omitted, fonts are shown distorted or pixels are magnified
Performance view: options: when switching tab close button is inoperable
Performance view: new incoming values are displayed only every 5min while 1min intervals are set
Object detail view: cluster partnership: attempts to issue the price ratio, which is unknown
Closing tab with+ click causes GUI to freeze & OutOfMemoryError
BVQ license window: high CPU load during visual display
Tab bar: icon is not updated after the first content modification
Tab bar: IndexOutOfBoundException when last tab is closed
Table view: open detail views cannot be opened anymore
Modifications in a GUI configuration file, regarding the scanner status checks are necessary when a topology scanner interval > 30min is configured. Please contact the BVQ support in this case.
|BVQ||BVQ Excel Export|
The Excel-format generated by BVQ Reporting is compatible with Excel Version 2003 and higher.
|BVQ||BVQ Scanner / Scheduler|
New features for scanner or scheduler which were unlocked by a new license file will take effect after about 2 hours or after restart of the corresponding services.
|IBM PMR 91765,033,724||SVC 22.214.171.124||Severe impairment of the VDisk latency by the SVC CLI command 'lsmdiskextent'. Seen at a cluster with 8 nodes (8G4) under 126.96.36.199. |
Starting with BVQ 2.4.3 this command is not used anymore. Instead command 'Isvdiskextent' is used.
Massive accumulation of reaching the SVC limit of 10 simultaneously competitive SSH sessions can trigger a warm restart of the node.
|BVQ-6610||BVQ VMware||Multiple mappings of a VM virtual disk (and/or of different snapshot versions of a single VM) to a VM virtual machine are not supported yet. |
This happens for example with flash copy managers like TSM4VE.
|BVQ-6623||BVQ VMware||Scan and representation snapshots and thier capacities are not yet supported by BVQ. The 'Used Capacity' of a VM Virtual disk corresponds currently to the 'Used Capacity' of the original VMDK file.|
Currently not supported VMware functions:
|IBM PMR ?||SVC Performance Stats|
Starting with version 188.8.131.52 SLAV, PLAV and DLAV refer to the consolidated backend IO sizes of up to 256k. But not the IO-counter for CTRM and CTP, which means that their quotient may be too low by a factor of 8. Therefore, the VDisk below cache metrics SLAV and PLAV cannot be used directly to assess the backend latencies. Also their transaction size is not the same of the backend (max 32k).
However, CTD was adapted to the consolidated IO sizes and provides continuously useful values.
|IBM PMR ?||SVC 7.3|
|Legacy VDisk cache statistics corrupted|
|IBM PMR ?||SVC 7.3|
|Node CPU core stats =0|
|IBM Defect 197194|
SVC 7.3 Performance Stats
|CPT, OPT, APT much too high|
|IBM PMR ?||SVC 7.4 Performace Stats||CPU core COMP (compression) statistics too high (factor ~100).|
DB schema changes
|184.108.40.206||SCHEMA_LEVEL||Increased schema level|
|Added port type to primary key due to V7000-300 models using identical WWPN for SAS and FC ports (BVQ-8589)|
See schema version 220.127.116.11
Requirements and restrictions
Requirements of the HW/SW environment
Minimum BVQ version required for an update
DB2 Advanced Workgroup Server Edition V10.5 FP5
|Java Runtime minimum version||18.104.22.168 (Java 1.8 not supported)|
Summary: update steps
- A new license file may be necessary to use certain new features.
- The GUI is possibly delivered with new predefined favorites. Only favorites contained in user favorites are preserved.
- Checking for correct function of the environment, in particular function of the scanner.
- Manual stop of BVQ Scanner and Alerting-service, Scheduler-services and also of all GUIs and reporting-clients
- Backup of BVQ installation directory and BVQ DB (e.g. DB offline backup with DB2 CMD)
- Starting again the BVQ Scanner
- Update BVQ Install Package Repository:
- DB Schema update
- BVQ GUI update
- BVQ Reporting Engine update
- BVQ Report Module update
- BVQ SVC Scanner update
- BVQ VMware Scanner update
- BVQ Scheduler update
- BVQ Alerting update
- Performing functional tests
- If necessary inform colleagues about the need to update locally installed BVQ GUI / report clients.