Store current job state directly in Riak
dmitrizagidulin opened this issue · 0 comments
dmitrizagidulin commented
For the same reasons as #121, we should store the state of Explorer long-running jobs in Riak, instead of in memory as is currently done.
This becomes important in a multi-Explorer-node setup (such as for Riak Mesos), so that more than one Explorer node doesn't accidentally start the same job when it could just wait for another node to finish.