Ram and Data usage for Couchbase Server 4.5 Mac OS vs Linux

Couchbase Server
,
1

Hi there,

I’m trying to track down an issue where a query seems to take a long time. While investigating the issue I noticed that I was unable to replicate the problem on my Mac whereas the reporters of the issue were using Linux. I then noticed that there seems to be a big resource discrepancy between MacOS and Linux. Here are 3 screenshots of a bucket we are using that are running on Linux servers:

filesystemqa
filesystemqa.png2084×716 136 KB

mark
mark.jpg1054×367 103 KB

longsheng
longsheng.png1086×427 53.7 KB

Here are 2 screenshots of a bucket running on MacOS servers:

filesystem-justin
filesystem-justin.png1045×357 52.8 KB

david
david.png1051×355 50.2 KB

I realize this is anecdotal and a super small sample size, but it looks like the usage on Linux is quite a bit higher than MacOS. This would help to explain what we’re seeing with the original query issue I’m looking into.

Has this higher usage on Linux systems been seen before?

2

The major difference is that Linux partitions a Bucket into 1024 vBuckets; whereas macOS partitions into 64. This has a noticeable affect on the per-Bucket overhead, especially when you have few (<1 million) items in a Bucket.

The reason for this is simply that macOS isn’t supported for production (it’s a development-only platform), and hence users will rarely use more than one node. The number of vBuckets essentially defines the maximum number of nodes a cluster can support (data is distributed across in whole vBuckets).

3

Sounds good, thanks!!