Limitations of config populate option of the installation toolkit
You can use the config populate option of the installation toolkit to populate the cluster definition file with the current cluster
state.
This functionality has the following limitations.
| Function | Description |
|---|---|
 Call home |
The config populate functionality does not support retrieving call home configuration
information. If call home is enabled in the cluster, use the ./spectrumscale callhome
enable, ./spectrumscale callhome config, and ./spectrumscale
callhome schedule commands to update the information in the cluster definition
file. |
| CES Groups | The config populate functionality does not detect CES groups. This might cause issues with CES stretch clusters |
| Clusters larger than 16 nodes | The config populate functionality does not support more than 16 nodes. For more information, see Limitations of the installation toolkit. |
| Custom GPFS™ profiles | The config populate functionality does not detect custom profiles. This might not be an issue in most scenarios except when you plan to use the config populate functionality to back up a cluster's configuration for a future new installation. |
| File audit logging |
The config populate functionality does not support retrieving file audit logging
configuration information. If file audit logging is enabled in the cluster, use the
./spectrumscale fileauditlogging enable and ./spectrumscale filesystem
modify --fileauditloggingenable
FileSystemName commands to update the information in the cluster definition
file. |
| FPO | The config populate functionality does not detect FPO NSD stanza information. However, the existing FPO NSD stanza information is not affected. |
| Greater than 3 GUI nodes | The config populate functionality adds in GUI nodes properly if the number of GUI nodes is less than or equal to 3. However, if there are more than 3 GUI nodes, the config populate functionality does not add them. This might leave you with GUI nodes that are not at the latest level of code that have to be manually upgraded, and other unexpected side effects. |
| Greater than 3 performance monitoring collector nodes | The config populate functionality does not support more than 3 performance monitoring collector nodes. If GUI nodes are detected, they are also assigned as the performance monitoring collector nodes. |
| Highly-available write cache (HAWC) | The config populate functionality does not detect HAWC information. However, the existing HAWC information is not affected. |
| Heterogeneous cluster | The config populate functionality cannot be used in a cluster that contains nodes with mixed CPU architectures or endianness. The installation toolkit can still be used in these type of clusters, but only on the subset of nodes that have the same CPU architecture and endianness. |
| IBM Spectrum Protect™ for Space Management and IBM Spectrum Archive™ Enterprise Edition (EE) | The config populate functionality does not detect IBM Spectrum Protect
for Space Management and IBM Spectrum Archive
Enterprise Edition (EE) information. The installation toolkit provides some DMAPI detection and gives recommendations on how to proceed in case issues are encountered. For example: DMAPI holds an FS open for one of these functions, thereby causing upgrades to fail. In this case, the installation toolkit shows a hint that identifies the node and that it might need to be rebooted to proceed with the upgrade. |
| IBM Spectrum
Scale release 4.2.0.2 |
In the IBM Spectrum
Scale
release 4.2.0.2, although the config populate functionality can be used to populate the cluster
definition file with the current cluster state, not all parameters are properly updated. For
example, CES IPs are not properly detected and you might need to add them manually by using the
following command: ./spectrumscale config protocols -eces_ip1,ces_ip2,.... Review your
cluster configuration after running ./spectrumscale config populate to ensure
that all desired values are populated. |
| Local read-only cache (LROC) | The config populate functionality does not detect LROC information. However, the existing LROC information is not affected. |
| Non-federated performance monitoring setup | The config populate functionality does not support non-federated performance monitoring setup. The installation toolkit converts all performance monitoring configuration to federated mode unless specified to ignore performance monitoring completely. |
| NSD SAN attachment |
The config populate functionality does not support retrieval of the NSD information in case
of an NSD SAN attachment. |
| Transparent cloud tiering | The config populate functionality does not detect Transparent cloud tiering related information. |
| Windows, AIX®, Z series | The config populate functionality does not support these operating systems. If any of these operating systems are detected on a node and if that node has the same CPU architecture and endianness as the node from which the installation toolkit is running, the config populate functionality shows a warning, skips the node, and continues populating information from other supported nodes. |