To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/CAM8WWJAsK0YaKnm-gKy_axL8d-YgbsK9eBEaWdh%2BH%3DXpribvHQ%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.
I could see swap vs no swap being an argument with no clear answer, because of varying circumstances and priorities.
The general question is, Is random access within a 2GB partition on a 200GB disk as fast as random access across the whole 200GB disk? I imagine various ways of sharding a virtual disk across the physical storage that answer this question differently, so what should we plan for?
To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/030e1adc-dc85-4bf4-83e9-eebee1c17583%40googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/030e1adc-dc85-4bf4-83e9-eebee1c17583%40googlegroups.com.
Well, if you don't know the priorities of his system, it's silly to say swap shouldn't be used. Some very smart people invented swap for some very good reasons, and we don't know whether his priorities align with those reasons.
The general question is important for lots of workloads. I wanna turn up MySQL with a smallish data set, but if the performance is gonna suck because a 2GB data set can't be performant, I need to look at alternative solutions.
Mine is really the same question as his, with a different application. Thoughts?
To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/3B0596D4-C0FD-4588-81D3-F3097DBB5CB6%40shortbus.org.