After creating (either create or batch create) a record for two tables, and then immediately: (1) querying those tables; (2) reading those records; (3) and deleting those records ---
I get these three errors in combination:
1) "The server response could not be downloaded at this time"
2) "Error: There was a problem getting a response from the server for cdb_batchCreateCloud request"
3) "Could not create record in cloud"
Any suggestion on how to avoid these errors?
SFO, as well as all the other regions, have not had any outages.
I suggest you try creating a record and attempting to query it in LCM. If it works there, it helps narrow down the issue.
I did a quick test and did not see any issues. Please check that you have authenticated your session before executing any other APIs. The other thing to check is that you have loaded all relevant tables.
I placed the errors in the order of how they happen. The short of it is that the client never heard back from the instance. This situation could occur if you were writing a large amount of data in a single call.
Here is a rundown on the process.
The client writes data to the cloud.
The client starts polling for a response from the cloud.
The cloud processes the transaction and writes a response for the client.
If the client times out, the interpretation would be that the cloud did not respond in time. Chances are very high that the cloud processed the transaction just fine, but we do not know for sure under this condition.
This condition would happen if one gave the instance a lot of data to transact. The "instance" did respond, but just not in time for the client to get it.
Options
You could adjust your time out value by calling this command: cdbi_setCloudTolerance "10000", "4000". The "10000" may need to be increased to "30000" (millisecs) or more. Some services will set that value to "90000" or more. Not all applications should use higher values. It depends on how responsive you want your app to be. For writes, I would not set it higher. I would probably ignore this particular error. The makers of this service should perhaps document this stuff better.
In the future, we have plans to stream data in more elegant ways that will prevent timeouts. I can't elaborate on this further at this time, as it has not been tested.
The instance may not have responded to the client with a positive or negative response.
• "Error: The server response could not be downloaded at this time."
Download Server Response code
This error happens when the client is expecting a message back from the cloud instance and the client times out waiting.
• "Error: There was a problem getting a response from the server for cdb_batchCreateCloud request"
Check Upload code
When data is written to the cloud. The client checks with the server for any errors. When the Download Server Response code gave up trying, this code will generate an error.
• "Could not write record to cloud."
cdb_batchCreateCloud
This error is near the end of the API. When the check upload code fails, the API responds with this error.
