Only _monitoring is being backed up after upgrade from 1.8 to 2.2 (then 2.4)


We performed an upgrade from 1.8 to 2.2 and then from 2.2 to 2.4 and we haven’t had a full backup since.

After the 2.2 to 2.4 upgrade we get _monitoring but nothing else and I cannot find why.
I’ve followed Back up data in InfluxDB | InfluxDB OSS 2.4 Documentation, using the admin token which returns fetching metadata then a few messages about shards.

Looking into the backups the only bucket backed up is _monitoring. If i try to specify a bucket it exits directly after fetching metadata.

I also had a look at the API that returns metadata and some buckets that have shards defined are in there and some are missing completely. These buckets can’t be found in the UI either but are listed in influx bucket ls.

I’m stumped so I’m hoping someone has any ideas?

Hello @rickard_eriksson,
Did you have buckets in your 1.x instance?
Yah that is odd. You should at a minimum be seeing the _task bucket as well.
I’m not sure what’s happening here. Let me loop someone in.

Hey @rickard_eriksson,

I’m a dev who has been investigating issues around upgrade from 1.8 → 2.x and have a few questions for you.

  1. Do you happen to have the original 1.8 directory still around and is there any chance you can privately share it with us? We haven’t been able to replicate on our own and are trying to get a reproduction case.
  2. Did you go from 1.8 → 2.2, use it for awhile and then go 2.2->2.4? Or did you basically just do 1.8->2.2->2.4? And if the first one, was everything working while you were on 2.2?
  3. Can you run influxd inspect report-tsm and verify that the cardinality is what you would expect?
  4. Can you run influxd inspect report-tsi on each of the databases/buckets listed by the above and see if it reports values you would expect?


We had no buckets in 1.x.

  1. Unfortunately we dropped the old data just a few weeks ago so it’s no longer available.
  2. We went 1.8 → 2.2 in may, used it until a couple of weeks ago then went to 2.4 hoping that it would solve the issues.
    3 + 4.

influxd inspect report-tsm

DB RP Shard File Series New (est) Min Time Max Time Load Time

Summary: Files: 0
Time Range: 2262-04-11T23:47:16.854775807Z - 1677-09-21T00:12:43.145224192Z
Duration: -2562047h47m16.854775808s

Total (est): 0
Completed in 136.794µs

So yeah, something feels wrong here :smiley:

If you are willing to do some live debugging/investigation, can you message me @Jeffrey Smith on the Community Slack and we can try to setup some time to look into this?

That’s definitely not the output I would have expected.