We need to validate that we can perform a no downtime upgrade. Maybe similarly to Elasticsearch's rolling upgrades as documented here.
Description
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | • Manybubbles | T90101 Confirm selection of BlazeGraph for wikidata query | |||
Resolved | • Manybubbles | T90103 BlazeGraph Finalization: Operational issues | |||
Resolved | • Manybubbles | T90445 BlazeGraph Finalization: Plan for no downtime upgrade |
Event Timeline
@Thompsonbry.systap, does BlazeGraph has a documented process for rolling upgrades in HA mode?
HA uses versioned RMI messages to avoid potential conflicts in rolling upgrades. To upgrade, simply shutdown a given HAJournalServer process. Redeploy the code. Restart the HAJournalServer. It will automatically resync and go live.
We try to avoid binary migration issues whenever possible and document any that exist at [1].
Help in documenting HA procedures would be welcome. We anticipate this being significantly easier as we move to standardize and simplify deployment models.
Bryan
Resolved. We'll want to perform rolling restarts in beta while we're performing updates to validate it. But we'll wait on that until we have the incremental updater.