Panic in jormungandr logs after stake pool retiremenet

[dkijania]

Describe the bug
There is an ugly panic in jormungandr after stake pool (the only one in blockchain) was retired:

"thread \'tokio-runtime-worker\' panicked at \'Distribution of rewards will not overflow: Delegation(NotFound($ hash_ty(0x38d18bcafb8ba65673dc00eb8ce7150ea3a7e61c24153dafa23cb437caa9b970)))\', jormungandr/src/blockchain/chain.rs:852:46"

Mandatory Information

1. jcli --full-version: newest master
2. jormungandr --full-version newest master

To Reproduce
Steps to reproduce the behavior:

1. start blockchain with single stake pool
2. retire it after some time
3. Observe issue

Expected behavior
node stats should be updated

Additional context
There is PR with automated test for this issue. #2145. Run test by:

 cd jormungandr-integration-tests
`cargo test test_pool_update`

[dkijania]

Panic is produced when we retire the only one stake pool in blockchain. Therefore it's only matter or better error reporting

[mzabaluev]

Is this reproducible after #2815?