Memory eating with 1.7.x

hi,

with 1.7.x I have massive problems with the memory. 12GB wasn’t enough anymore and my VM has now 32GB … not enough OOMs every few hours.
I’m not alone with it: Forum and also on Guthub from InfluxDB with a search for memory brings a lot of them. I switched to TSI1, but it was useless.
What is interesting: If I take a look on “show shard groups;” and search for the IDs on the disk … there are lot, which are not there. No idea, if it is good, or bad. I assume the 2nd.

Any suggestions ? I think, Influx takes now only care of the influxDB2.x and the cloud version. :-/

Did you just change the setting to ts1 in the ini file or did you also convert your data to ts1?
If you didnt create the index files, all the old data is still the same.

hi,

sorry, I’ve totally forgotten to answer … jepp, data also migrated. Instead of just let kill the process from the kernel and let Systemd let start the process again, I do a restart every hour for the Influxdb. If you ask me … I think, they focus only on 2.0 and the cloud version.
We get an new ISCSI SSD storage the next weeks and I hope, that it will be faster to write the data down. Maybe the problem disappears. If not … I’m thinking really to try out Prometheus.