- 最新
- 最多得票
- 最多評論
Hi egao91,
Is your issue with the OOM errors or Redshift SELECT and UPDATE statement execution time?
Clearly the OOM error is an issue. I'd recommend you direct that to AWS Support if you have a support plan higher than basic forum based support.
WRT your statement response times, Redshift, as an analytic column-store SQL database, is not built to perform a lot of small narrow queries (like your SELECT) or do a lot of small UPDATE statements (you didn't provide your UPDATE statement so I'm assuming a single row update) like an OLTP SQL database with row-store storage is built to do. Instead small narrow query and DML statement response time, it's tuned for overall large complex query workload throughput and fast loads (COPY command) of huge amounts of data.
I hope this helps you at least a little bit,
-Kurt
Thanks Kurt. My first issue is the oom error. I tried to limit the connection pool to 10. and then it happened rarely. I don't know what's cause the OOM. I checked the WLM and didn't find any helpful messages. because the table is small enough. and the database may provide some mechanism to avoid oom (eg using a cache file ) for this kind of simple select and update.
after I limited the connection pool to 10 (a really small pool), the query (a very selective query with where conditions) cost about 10 seconds. and update (also similar with select) cost 700 ms. but when I stopped some tasks (which will use redshift), the time is recovered to a reasonable value about 70ms.
Edited by: egao91 on Jan 31, 2021 10:26 PM
Hi egao91,
Glad to see you're making some progress on your own. I think that's all useful information for AWS Redshift Support to help them diagnose the 2 issues. I think you need their help from here. I don't know of anyone the Redshift customer community here that help diagnose this further and get a fix.
Regards,
-Kurt
相關內容
- AWS 官方已更新 1 年前