An inhomogeneous management cluster is a management cluster consisting of an SE server with all units (MU, HNC, SU x86) with version V6.2 and another SE server with all units (MU, HNC, SU x86) with version V6.3 This situation can occur if the two servers of a management cluster are upgraded in two steps.
In general it can be said that new functions should not be used for units with version V6.2, even if the new functions are displayed for them.
This chapter deals with specific topics and limitations that must be considered in an inhomogeneous management cluster.
Unit states
New system states for units may occur in V6.3:
- If the resource monitor (ResMon) detects a resource problem the system status for this unit is set to WARNING.
- If an AU has been registered only for HW monitoring its system status is NOT_MONITORED.
These states have not been foreseen in older versions resp. in V6.2.
Therefore on V6.2 MUs there will be the following restrictions:
- Units and systems overview windows
- Units: The AUs with system status NOT_MONITORED and the units with system status WARNING will be displayed using the icon generally used for UNKNOWN / NOT_ACCESSIBLE.
- Systems: The systems associated to NOT_MONITORED AUs are Native-AU systems and will be displayed using the icon generally used for UNKNOWN / NOT_ACCESSIBLE.
- In the tables the corresponding filter values will be missing.
- The SEM menus for these units will not be be displayed or be restricted due to the unknown status.
(hardware, devices, systems) - SEM Dashboard: Due to the unknown value the counting in the dashboard is not correct:
- In the tile view the NOT_MONITORED AUs are not counted so that the counter of all units is not the sum of the units for the single states.
- In the list view the NOT_MONITORED AUs are missing and the sum per line doesn't fit.
- In the list view the Native-AU systems of the NOT_MONITORED AUs are missing and the sum per line doesn't fit.
Audit logging entries
Regarding login and logout the audit logging has been enlarged in V6.3 with new entry types.
In V6.2 these entry types are not known so that internal values are displayed like "login_successful" or "logout" in the Type column.
In V6.3 also proper filter values are missing for these entry types.
Alarm management and component-specific notifying
- Mail and trap receivers created on V6.2 MUs (or restored by CSR from V6.2) will be visualized and handled on V6.3 MUs as being defined for Component = ANY and cannot be distinguished from a new entry with explicit specification of Component = ANY. (In this situation duplicate entries may occur and alarms are sent twice. It makes no difference which entry will be deleted.)
In case they have to be made component-specific they must be deleted and created again. - Mail and trap receivers created on V6.3 MUs with the same addressee but with different components will be displayed as duplicates on V6.2 MUs.
If the cluster manager is a V6.2 MU alarms for these addressees are sent multiple and regardless of the components specified.
Creating IORSF files and transferring them to the SVP
The generation of IORSF files and their transfer to the SVP always takes place on the MU with the status ACTIVE in relation to the SVP console of the SU /390 of the respective server.
Restriction:
The new functions are not available on MU with version V6.2.
If this function is called for such a MU, this leads to an error: "Not Found. The requested URL .... was not found on this server".
Update IORSF file list
- In large configurations with many BS2000 devices the update of the IORSF file list may last very long and may exceed the timeout value set for the dialog of SEM. In this case the dialog will be aborted with an error message.
This happens on V6.2 MUs where the dialog is aborted with the message "The LAN connection was disrupted and the result of the action could not be determined.".
Nevertheless, the update of the IORSF file list is continued in the background and comes to a successful end.
On V6.3 MUs this does not happen.
Therefore, the user is advised to call this SEM dialog always on a V6.3 MU, even if the concerned MUs are V6.2 MUs. - When a management cluster is established the IORSF file and device lists for one of the two servers is lost.
Please check the IORSF file and device lists on both servers and update where necessary.
Set the name of the SU /390
Restriction:
The name of an SU /390 belonging to a server with units of version V6.2 can be set on a V6.3 MU, but is not recommended.
The name will be displayed on the MUs with version V6.3 only for a short time after setting it and will disappear soon.
Call of the SVP console
Restriction:
In an inhomogeneous management cluster the SVP console can only be accessed using the SE Manager of an MU of the same SE server the SU /390 is part of.
(-> Use a V6.2 MU for SVP console on V6.2 server and use a V6.3 MU for SVP console on V6.3 server!)
Call of the shadow terminal
Restriction:
In an inhomogeneous management cluster the shadow terminal on an MU can only be accessed using the SE Manager of an MU of the same SE server the addressed MU is part of.
(-> Use a V6.2 MU for shadow terminal on V6.2 server and use a V6.3 MU for shadow terminal on V6.3 server!)
Installation of hot fixes
The format of the hot fixes has been changed to M2000/HNC/X2000 V6.3.
Old format example (<= V6.2): MV6.2A0501.A099999990-H05
New format example (>= V6.3): MV6.3A0456H012
Meaning of MV6.3A0456H012:
- The 2nd Increment of the Hot Fix 01 for M2000 V6.3A0456.
- A hot fix generally summarizes the corrections for several problems (PMs).
Restriction:
In an inhomogeneous management cluster, hot fixes can only be uploaded via the SE Manager of an MU of the same SE server to which the target unit belongs. (-> Use V6.2 MUs to upload V6.2 HFs and V6.3 MUs to upload V6.3 HFs!)
New main windows
In V6.3 there are three new main windows in comparison with V6.2:
- Systems -> <SU /390> -> SVP console
- Hardware -> Units -> Update overview
- Service -> Information
If one changes the MU (via the header functionality) from a V6.3 to a V6.2 MU with one of these windows open, the corresponding window will not be found and "Not Found / The requested URL … was not found on this server" will be displayed.
Situational dependencies (Cluster Manager on V6.2 MU, MU powered off, …)
In case that the Cluster Manager runs on a V6.2 MU the so-called global data are enquired on that MU.
This means that some data introduced in V6.3 will not be available and on the V6.3 side resp. for V6.3 units there will be some restrictions.
- System status WARNING for units:
- Since the resource monitor is not called there will be no status WARNING.
- The event “State of unit '<hostname>' changed from '<oldstate>' to 'WARNING'” will not be created.
- IP networks
- The causes for the states different from NORMAL are not detected and will not be displayed.
- Also the warning icon in the dashboard will not be displayed.
In case that the V6.3 MUs are powered off and the enquire processes run on the V6.2 MUs there will be some additional restrictions:
- System NOT_MONITORED for units:
- This system status is not known in V6.2. Instead status UNCONFIGURED will be displayed.
CLI remoteCommand
Restriction:
In an inhomogeneous management cluster remoteCommand should only be used within one server, i.e. MU and target unit (currently SU x86) should be in the same server. (-> Use V6.2 MUs for V6.2 SUs and V6.3 MUs for V6.3 SUs for remoteCommand calls!