4

we are getting started with a mongo cluster. The best practices docs mention that the ideal storage capacity per node is ~5TB. Given we can get ~4TB from a single SSD, we are planning to go with it instead of HDD. Is there any gotchas that we need to be aware of running 1xSSD per node?

Thanks in advance.

Dan
  • 51
  • 3

1 Answers1

-4

You mean outside the obvious?

  • Failure means node down. No redundancy at this level (which may be ok if you have lot of higher up redundancy).
  • You are limited to the IO performance of one SSD. If that one is SATA that is QUITE low compared to a raid of smaller ones or an M.2 SSD.
TomTom
  • 51,649
  • 7
  • 54
  • 136