Nodes can't resume after out-of-disk failure: Assert_failure vendors/irmin-pack/pack.ml:270:6

Question

• 2 Nodes on same host : one archive-mode mainnet and a babylonnet baker
• both failed when machine ran out of disk space
• cleaned-up space but restarting them shows:

Jan 30 13:43:38 - validator.block: Validation of block BKjHmnynhftFBaG98rjG32UNFUvC5uRhJh6FbnKmwWT51L8MEWW failed
Jan 30 13:43:38 - validator.block: Request pushed on 2020-01-30T18:43:38-00:00, treated in 21us, completed in 333ms , 
Jan 30 13:43:38 - validator.block: "Assert_failure vendors/irmin-pack/pack.ml:270:6"
Jan 30 13:43:38 - node.validator.bootstrap_pipeline: Unexpected error (validator):
Jan 30 13:43:38 - node.validator.bootstrap_pipeline: Error:
Jan 30 13:43:38 - node.validator.bootstrap_pipeline:   "Assert_failure vendors/irmin-pack/pack.ml:270:6"
Jan 30 13:43:38 - node.validator.bootstrap_pipeline: 
Jan 30 13:43:38 - validator.peer: Worker crashed [validator-peer_240]:

For both nodes tezos-admin-client list rejected blocks says:

No invalid blocks.

And tezos-admin-client unmark all invalid blocks does nothing.

 $ uname -srvmpio
Darwin 18.7.0 Darwin Kernel Version 18.7.0: Sun Dec  1 18:59:03 PST 2019; root:xnu-4903.278.19~1/RELEASE_X86_64 x86_64 i386 Macmini8,1 Darwin

It's the multinetwork-node from a while ago:

 $ tezos-node --version
307cae9e (2019-10-11 15:45:11 +0200)

PS: I had seen somewhere dune exec -- vendors/index/fix.exe ..., there does not seem to be such an executable definition in vendors/index/..