seaweedfs filer panic after server crash and restart

[tianjipeng]

Sponsors SeaweedFS via Patreon https://www.patreon.com/seaweedfs
Report issues here. Ask questions here https://stackoverflow.com/questions/tagged/seaweedfs

Describe the bug
We use leveldb2 as filer storage, after server crash and restart, filer will always panic when start it.

System Setup

• command line: weed filer
• OS version: centos7.4
• output of weed version: 1.61
• content of filer.toml:
  [leveldb2] enabled = true dir = "/filer" # directory to store level db files

Expected behavior
filer run normally

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context
I find the root cause that return the wrong err in weed/filer2/leveldb2/leveldb2_store.go :

Meanwhile, leveldb2 provide Recover function to repaire corruted data，so if dbErr is ErrCorrupted，we should call Recover function.

[chrislusf]

Thanks a lot for the debugging! Good to learn this Recover feature!