MPDS FAQs
This section contains frequently asked questions regarding the MPDS service. The FAQ is divided into the following sections:
Backup and Restore
Q: How are the backups validated?
A: Backups are validated with pgbackrest and periodic CIS reports.
Q: How are missing backups detected?
A: Backup jobs are monitored with pgoperate and Icinga.
Q: Can the customer initiate a database backup manually or on-demand?
A: No.
Backups are fully automated.
Q: On what basis were decided the min cpu and max cpu numbers?
A: Minimum and maximum CPU numbers are calculated as follows:
Minimum: the minimal allocation is one physical CPU core. Therefore, the minimum order size is 2 vCPU (=2*0.5 = 1 Intel x86 Core).
Maximum: the maximum vCPU allocation per container is limited by the physical memory capacity of the compute node (768 GB physical RAM - 64GB host RAM = 704 GB; 704:8 = 88 vCPU)
Q: How do I restore a database?
A: Database restore is a day-2 action on instance level. Check the restore PDB section.
Operation and Support
Q: Container restart - when and how?
A: In general, there is no resource action available to restart the container on-demand. The container will automatically restart in two cases:
When changing the number of vCPUs. This is a modify action. It requires a re-start of the container, because not only more (or less) vCPUs are allocated, but also the RAM allocation needs to be adapted.
When enabling or disabling database backup. This is a modify action. It requires a restart of the container, because the archive mode of the database needs to be enabled/disabled.
See also Day-2 Action Overview.
Q: How can I monitor database activity?
A: Database activity can be retrieved from postgreSQL statistics views.
Q: Can you monitor the VM utilization of your Container?
A: Yes.
The utilization of the VM can be visualized whithin the vRA console. Note that this feature is only available for customers with a swisscom tenant.
Q: Which IP ports will be assigned to MPDS instances?
A: IP Ports which will be assigned to MPDS instances start from 49152 up to 65535 (within the same tenant).
First provisioned MPDS instance will get port 49152, second 49153, ...
Ports from deleted instances will be reassigned to new instances.
Security
No security related FAQ so far.
SW Lifecycle and Patching
Q: How often will the HW stack be patched?
A: The PCA will be patched once a year. Security patches will be installed according their criticality. If required and/or recommended by the vendor support (e.g. because of know bugs) patches will be applied immediately.
Q: What PostgreSQL releases are available on the MPDS platform?
A: See PostgreSQL Releases for details.
Q: How often are existing deployments patched?
A: Existing deployments (container, database instances) will be patched on service request only. Emergency security patches can be installed without the explicit consent of the customer. Nevertheless: customers will be informed as early as possible.
ADVANCED Plan
Q: How can I monitor the Primary/Standby configuration?
A: Replication activity can be retrieved by SQL on Primary and Standby instance from statistics views.
Q: Is it possible to switchover a database on demand?
A: No.
There is no edit instance action available to switchover a database. A switchover can be requested by a service request.
Q: Can I revert an ADVANCED configuration to a BASIC configuration?
A: No.
You have to provision a new BASIC instance and export/import the data. Support will be provided on-demand.
Q: Can I use the standby database read-only?
A: Yes.
It's possible to query the standby db (hot_standby is default set to ON).