Hi - we’re experiencing an issue where Influxdb (0.8.8) crashes periodically on summarization. I believe this is due to having too many series in our database due to fanouts and continuous queries. We have over 2,000,000 series. To pile on that problem, we are using the default shard space - retention inf, duration 7 days.
Clearly, we could have better planned our DB architecture. I’d like to be able drop over million series however, I’m unable to even drop a single series without crashing the instance. I’ve also tried updating the shard retention to 30 days to see if that makes the deletion process easier but it doesn’t seem to make a difference. Any advice on the best way to move forward?
Some thoughts are:
- Start fresh on a new database; this is the least ideal as we wouldn’t migrate data
- Update shard space retention to 30 days and reduce the amount of data we’re writing plus eliminate continuous queries and fanout. After the 30 days has passed drop all data older than 30 days.