Metrics errors on a clean node startup

[LesnyRumcajs]

Describe the bug

A fresh instance of Forest, i.e. with no ~/.forest (or rather config.data_dir + /db) is triggering an error in metrics. It shouldn't be a big issue but it's not nice for a new user to see errors on a clean instance.

To Reproduce
Steps to reproduce the behavior:

1. Remove your rocksdb directory, e.g. $HOME/.forest/db
2. Run forest as usual
3. See the error as below error

Log output

❯ target/release/forest --target-peer-count 50 --encrypt-keystore false --import-snapshot ~/Downloads/minimal_finality_stateroots_latest.car 2>&1 | tee forest.log                                                                                                                                                            
2022-03-29T11:07:56.384Z WARN  forest::cli > No configuration found! Using default                                                                                                                                                                                                                                           
2022-03-29T11:07:56.384Z INFO  forest::daemon > Starting Forest daemon, version v0.2.2/unstable/mainnet/5f0c7f0f                                                                                                                                                                                                             
2022-03-29T11:07:56.384Z INFO  forest_libp2p::service > Recovered libp2p keypair from "/home/rumcajs-work/.forest/libp2p/keypair"                                                                                                                                                                                            
2022-03-29T11:07:56.384Z WARN  forest::daemon         > Warning: Keystore encryption disabled!                                                                                                                                                                                                                               
Admin token: eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJBbGxvdyI6WyJyZWFkIiwid3JpdGUiLCJzaWduIiwiYWRtaW4iXX0.4t86AoJCoAxEMPm9w2RNrmM09aJCR789SLVfkLUkclU                                                                                                                                                                         
2022-03-29T11:07:56.384Z INFO  metrics                > Prometheus server started at 127.0.0.1:6116                                                                                                                                                                                                                          
2022-03-29T11:07:57.629Z INFO  genesis                > Initialized genesis: BlockHeader: Cid(bafy2bzacecnamqgqmifpluoeldx7zzglxcljo6oja4vrmtj7432rphldpdmm2)                                                                                                                                                                
2022-03-29T11:07:57.630Z INFO  forest::daemon         > Using network :: testnetnet                                                                                                                                                                                                                                          
2022-03-29T11:07:57.630Z INFO  genesis                > Importing chain from snapshot                                                                                                                                                                                                                                        
2022-03-29T11:07:57.630Z INFO  genesis                > Reading file...                                                                                                                                                                                                                                                      
Scanning blockchain 4840 / 1534080 [>--------------------] 0.32 % 100.07/s 255m  2022-03-29T11:21:55.587Z ERROR metrics::db            > Calculating DB size for metrics failed: Error { kind: NotFound, message: "No such file or directory (os error 2)" }          

Expected behaviour

No errors from metrics on a fresh instance of Forest.

Screenshots

Environment (please complete the following information):

• OS: Linux
• Rust version(e.g. rustc --version) 1.59.0
• Branch/commit e2accbe

Other information and links

We should probably create the correct structure before initialising Prometheus though maybe the problem is more complex than that.

[lemmih]

Wasn't this fixed by #1582? @tyshko5, maybe you could look into this.

[LesnyRumcajs]

Ya, I thought it would be closed but I'm pretty sure I saw it recently. Let me double-check to confirm, perhaps I was running an old build.

[lemmih]

Maybe the metrics try to query the DB size before the DB is created.

[LesnyRumcajs]

Ya, can't see it anymore. I'll reopen it again with some strong proof if it happens again. :)