You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm attempting to backfill Prometheus data into Mimir, and for that I needed to use the splitblocks tool to split my large blocks into smaller ones that only span a day.
However, splitblocks tells me the following for each block it processes:
level=info ts=2025-01-25T02:01:13.3734175Z block=01HK2ZFKHRNYFHRN40NZJ403M6 block_min_time=2023-12-31T17:22:34.386Z block_max_time=2024-01-01T12:00:00Z
level=debug ts=2025-01-25T02:01:13.4089947Z block=01HK2ZFKHRNYFHRN40NZJ403M6 msg="not downloading again because a provided path matches this one" file=01HK2ZFKHRNYFHRN40NZJ403M6/meta.json
level=info ts=2025-01-25T02:01:13.4104421Z block=01HK2ZFKHRNYFHRN40NZJ403M6 msg="splitting block" minTime=2023-12-31T17:22:34.386Z maxTime=2024-01-01T00:00:00Z
level=error ts=2025-01-25T02:01:13.4122069Z block=01HK2ZFKHRNYFHRN40NZJ403M6 msg="failed to split block" err="failed injecting meta for split: invalid argument"
failed injecting meta for split: invalid argument
First I was worried that maybe I screwed up the data while moving, but then I decided to give it to a prometheus server, and it can read the metrics I checked fine. Checking was done by running a query with its own web interface, and the graph view was showing information I expected.
How to reproduce it?
Build splitblocks from source with go 1.23.5
Run splitblocks as ./splitblocks --backend filesystem --filesystem.dir data --output.dir prometheus_stats_split --block-concurrency 1
See the error messages as processing goes. For most blocks it takes a longer time
What did you think would happen?
I expected that splitblocks successfully splits the blocks, so I can continue with backfilling it to Mimir
What was your environment?
I was running splitblocks and the test promeheus server on my workstation, in WSL.
The prometheus data was generated over time in a debian system, where prometheus was running in docker.
Any additional context to share?
Could you please release binaries of splitblocks for future releases?
The text was updated successfully, but these errors were encountered:
What is the bug?
I'm attempting to backfill Prometheus data into Mimir, and for that I needed to use the splitblocks tool to split my large blocks into smaller ones that only span a day.
However, splitblocks tells me the following for each block it processes:
First I was worried that maybe I screwed up the data while moving, but then I decided to give it to a prometheus server, and it can read the metrics I checked fine. Checking was done by running a query with its own web interface, and the graph view was showing information I expected.
How to reproduce it?
./splitblocks --backend filesystem --filesystem.dir data --output.dir prometheus_stats_split --block-concurrency 1
What did you think would happen?
I expected that splitblocks successfully splits the blocks, so I can continue with backfilling it to Mimir
What was your environment?
I was running splitblocks and the test promeheus server on my workstation, in WSL.
The prometheus data was generated over time in a debian system, where prometheus was running in docker.
Any additional context to share?
Could you please release binaries of splitblocks for future releases?
The text was updated successfully, but these errors were encountered: