I have a big issue since yesterday. One DevOps put MyISAM database into Percona XtraDB Cluster and after that all nodes went to -1 state seqno (in grastate.dat) but besides that cluster works fine - is synced and no problems at all. We’ve removed that MyISAM DB and try lot of things like bootstraping cluster again even deleting whole mysql dir and putting back from xtrabackup. Also I tried mysld_safe --wsrep-recover (something like that) - it founds last seqno but after bootstraping node it goes to -1 again.
A sequence number of -1 on grastate.dat is a runtime value, on proper shutdown this will be replaced with the correct persistent sequence number. If the server crashes or dies for some reason, this -1 will indicate that the node will have to consult the other nodes for state i.e. does it need IST or SST.
In short, if your node is running, -1 is fine, if its not running something bad happened