Why is postgres slow when data is stored in a Volume as opposed to the local filing system?

Hi, I’m using a postgres database deployed as a docker container on a $15/mo Droplet (with swap). I’ve replicated the container for an A vs. B ‘Volume vs local filing system’ speed test.

Where the database files are : [A] on local machine file system/disk <---- takes 10s [B] on a 10GB Volume <— takes 34s

A query on a 7 million row table (see below) takes 10 secs on [A] and 34 secs on [B]

SELECT * FROM ( SELECT val1,val2, cdatetime, a, b, c, d FROM events_table WHERE cdatetime >= ‘2022-01-16T12:26’ AND cdatetime < ‘2022-01-17T15:26’ ORDER BY cdatetime DESC ) as sub_sigevent WHERE sub_sigevent.d = 59 AND sub_sigevent.a = 6 AND sub_sigevent.b = 3 AND sub_sigevent.c = 7

As Volumes are networked memory in effect, I assume this is caused by the slow connection to the Volume compared to local filing system. So it would appear there’s no way round this, other than to try a time series specific data base such as ClickHouse (OSS version of InfluxDB) that scales better.