GUI fails to start

This issue is primarily because of the database issue. In ideal scenarios, the service script automatically initializes and starts postgreSQL. However, in rare cases, the database might be either inconsistent or corrupted.

If the postgreSQL database is corrupted, it might be because of the following reasons:
  • The additional (non-distro) postgreSQL package is installed and it occupies the port 5432.
  • Details that are stored in the /etc/hosts file are corrupted so the "localhost" is not listed as the first item for the IP127.0.0.1.
  • An incompatible schema exists in the database from a previous release.
  • The GUI version is downgraded to an older version.
If the GUI logs show any of the database errors, try the following steps:
  1. Issue systemctl stop gpfsgui to stop GUI services.
  2. Issue 'su postgres -c 'psql -d postgres -c "DROP SCHEMA FSCC CASCADE"''.
  3. If the previous step does not help, issue 'rm -rf /var/lib/pgsql/data'.
  4. Issue systemctl start gpfsgui to start GUI.

If the problem still persists, it might be because of a corrupted GUI installation, missing GUI dependency, or some other unknown issue. In this scenario, you can remove and reinstall the GUI rpm. For more information on how to install and uninstall GUI rpms, see Manually installing IBM Storage Scale management GUI .

You can collect the logs that are available in the /var/log/cnlog/mgtsrv folder to investigate further. You can also use the gpfs.snap command as shown in the following example to collect logs and dumps in case of a GUI issue: 
gpfs.snap -N GUI_MGMT_SERVERS
Collecting logs and dumps through the gpfs.snap command also collects the GPFS logs. So, manually getting the logs from the folder /var/log/cnlog/mgtsrv is quicker and provides only the required data that is required to search for the details of the GUI issue.