Timeouts using get or get_any_replica BUT not using admin UI

I am getting <class ‘couchbase.exceptions.UnAmbiguousTimeoutException’> exception when trying to get binary data from the cluster base on document key. I have tried both get and get_any_replica. BUT when I enter that key into admin ui documents I am able to get the document. What should I do ? Is the REST endpoint that admin UI is using ready for high traffic and can I used it ? If I am using get_any_replica does it matter which server I am connected to ? does it matter with get?

I have also observed this behavior: when using get_multi with a list of for example 100 keys I am getting only 66 documents. When using n1ql query with meta().id in $keys with the same list of keys I am getting 99 rows. I have switched to n1ql query for that part but for binary data I need to use get

Language: Python
SDK version: 4.1.4
Cluster: 6.6 community

The Query API has a default timeout of 75 seconds. The KV API (get, get_multi, get_any_replica) has a default timeout of 2.5 seconds. Maybe trying specifying a longer timeout in the get_any_replica and get_multi calls.

I have set timeout for get to 20 seconds and to get_any_replica to 120 seconds. No luck.

Hi @orson – Could you provide DEBUG or TRACE level logs from what you are seeing? There might be something useful in the logs to help diagnose what might be happening. Also, sample code can be useful for diagnosing.

See logging docs here. Quickest path might be simply setting the PYCBC_LOG_LEVEL environment variable. Example: PYCBC_LOG_LEVEL=trace python <your script here>.py

Maybe the KV port is not accessible. Try running SDK Doctor against your server.
Or try accessing documents from Web Console → Buckets…

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.