Timeouts on certain document queries

I have a “flagged” view that emits documents with a “f” property equal to 1. I’m simply using get_multi and set_multi to process 500 documents at a time from this view (code sets the “f” property to 0, adds the document to a dict, calls set_multi with this dict when all documents are processed).

This runs great locally, but only worked for the first request against my production Couchbase cluster. Now every request for a “flagged” document times out, regardless of whether it is a get_multi request or a single get request. Requests for other documents work. In the web console I can see these documents that requests are timing out for; they definitely exist. Has this flagged view and its related documents become locked or gone into a state that would prevent querying these documents?

It’s really odd that requests for these specific documents time out. The timeout should mainly relate to the socket for the server rather than their presence inside views.

Can you tell me what server version and client version you’re using? If Client is 1.2.0, are you using twisted/gevent?

I’m not using Twisted or GEvent. I was on version 1.0 and just upgraded to 1.2 and tried this again, still no luck. I’ve run these queries from two different machines and both timeout for existing documents related to this flagged view.

This definitely appears to be something going on with the cluster itself. All server nodes are up and appear to be good. I’ve run these queries against different server node IPs as well as deleted the flagged view completely and re-created it, but this issue remains.