capture errored filesystems to provide to tezos devs

[orcutt989]

Errors such as

│ init-tezos-filesystem Error:                                                                                                                                                      ││ init-tezos-filesystem   "Assert_failure src/log_file.ml:73:4"                                                                                                                     │
│ init-tezos-filesystem tezos-node: Error:                                                                                                                                          ││ init-tezos-filesystem               "Assert_failure src/log_file.ml:73:4"

Occur frequently during storage initialization. Right now we just time the initalization to 2 minutes and toss the job because Tezos does not exit when it encounters such an error.

The Tezos Devs are interested in looking at these errors, but they need a tarballed filesystem to do this.

SInce we toss the job, we lose the filesystem. We will need to implement some way to tarball and label the artifact as an error with the specific error message to provide to them.