Hopefully somewhat simple question: influxdb is just extremely slow to start on my RaspberryPi I’ll enter the command influx and literally just sit around for a good 10 minutes until I get the command line prompt.
This has a cascade on effect during reboot when none of my python scripts that write to influxdb or my Grafana visualiser can load properly and basically means I just have to sit around for a bit after each reboot.
Surely this user error on my part somewhere along the line, but anyone have any ideas?
I’ve tried re-installing Influx twice, same result both times.
By default, at startup, influx is creating memory indexes for series.
Like a PI doesn’t have a lot of RAM, the result is that it’s swapping (more than) a lot at startup.
In /etc/influxdb/influxdb.conf try to use “tsi1” instead of “inmem” indexes:
# The type of shard index to use for new shards. The default is an in-memory index that is
# recreated at startup. A value of "tsi1" will use a disk based index that supports higher
# cardinality datasets.
# index-version = "inmem"
index-version = "tsi1"
Eventually you can also try to reduce memory cache size
# CacheMaxMemorySize is the maximum size a shard's cache can
# reach before it starts rejecting writes.
# Valid size suffixes are k, m, or g (case insensitive, 1024 = 1k).
# Values without a size suffix are in bytes.
cache-max-memory-size = "384m"
#cache-max-memory-size = "1g"
On my PI Zero it fix the issue. It’s still a bit slow (2-3 min) to starts but it’s starting!
BTW, using tsi1 will create disk indexes, so it will generate more I/O on disk, so take care of your SD life.
