Cluster add --wait fails


After adding a third member to my IPFS cluster, adding with wait flag has started to fail like:

 ${IPFS_CLUSTER_CTL} add --wait --quieter --format car ${CONTENT_CAR} 
panic: interface conversion: multiaddr.Multiaddr is nil, not *multiaddr.multiaddr
goroutine 23 [running]:*multiaddr).UnmarshalJSON(0xc000410360, {0xc0002b64fc, 0x5f, 0x304}) +0xe8*Multiaddr).UnmarshalJSON(0xc001234850, {0xc0002b64fc, 0x5f, 0x304}) +0x74
encoding/json.(*decodeState).literalStore(0xc000430630, {0xc0002b64fc, 0x5f, 0x304}, {0xe9dca0?, 0xc001234850?, 0x197?}, 0x0)
	encoding/json/decode.go:871 +0x1f2c

No idea why....

I have verified that each peer in the cluster see the other 2, so, I assume the cluster is well formed.

What could be going on?

This message is also taking place with other cluster operations, such as adding a pin.

looks like upgrading ipfs-cluster-ctl from 1.0.2 to 1.0.5 resolves the issue.

If you were using Kubo >= v0.18, it was probably ipfs v0.18: "ipfs does not seem to be available after x retries" · Issue #1835 · ipfs-cluster/ipfs-cluster · GitHub