Page created: 4 Feb 2020 |
Page updated: 22 Jul 2020
If a large number of client requests are receiving error responses, then view the logs/failed-ops log, which is an access log for only failed operations. The potential reasons for the error responses include the following:
- If clients are requesting operations that legitimately should fail (for example, they are targeting entries that do not exist, are attempting to update entries in a way that would violate the server schema, or are performing some other type of inappropriate operation), then the problem is likely with the client and not the server.
- If a portion of the Directory Proxy Server data is unavailable (for example, because an online LDIF import or restore is in progress), then operations targeting that data will fail. Those problems will be resolved when the backend containing that data is brought back online. During the outage, it might be desirable to update proxies or load balancers or both to route requests away from the affected server. As of Directory Proxy Server version 3.1 or later, the Directory Proxy Server will indicate that it is in a degraded status and the Directory Proxy Server will route around it.
- If the Directory Proxy Server work queue is configured with a maximum capacity and that capacity has been reached, then the server begins rejecting all new requests until space is available in the work queue. In this case, it might be necessary to alter the server configuration or the client requests or both, so that they can be processed more efficiently, or it might be necessary to add additional server instances to handle some of the workload.
- If an internal error occurs within the server while processing a client request, then the server terminates the connection to the client and logs a message about the problem that occurred. This should not happen under normal circumstances, so you will need to work with your authorized support provider to diagnose and correct the problem.
- If a problem is encountered while interacting with the underlying database (for example, an attempt to read from or write to disk failed because of a disk problem or lack of available disk space), then it can begin returning errors for all attempts to interact with the database until the backend is closed and re-opened and the database has been given a change to recover itself. In these cases, the je.info.* file in the database directory should provide information about the nature of the problem.