CODEWITHSUNDEEP
mongodbswisscomdev
minettiandrea
minettiandrea

Reputation: 61

MongoDB behavior on simple instance fail

I've seen that the standard MongoDB service on Swisscom Application Cloud is not in HA, so which level of service we should expect in case of an instance failure:

Upvotes: 2

Views: 90

Answers (1)

Lafunamor
Lafunamor

Reputation: 753

For the public offering Swisscom does not provide any guarantees concerning SLAs. That is including uptime. In case of the failure that is is within the responsibility of Swisscom, they are able detect the failure and start the recovery procedure in reasonable time. Swisscom strongly recommends using the HA services. I suggest reading through the terms of Swisscom Application Cloud: http://developer.swisscom.com/terms/

According to my experience the services restart quite fast, so automatic reconnect should be sufficient.

Generally the MongoDB Docs answer a lot of your questions: If a single instance MongoDB service fails, it will automatically recover all information stored on the disk as long as no corruption of the data happened. Have a look at: https://docs.mongodb.com/manual/core/journaling/

I also strongly suggest you perform regular backups trough the Application Clouds Portal. So In case of disk corruption you can restore your data.

TL:DR

  • The service will restart on a crash.
  • Restart time varies as it heavily depends on why and how it was triggered. Generally it takes longer if MongoDB has to restore data and reapply the journal.
  • Always enable automatic reconnect for your database backend.
  • Additionally: regularly backup your database.

Upvotes: 2

Related Questions