Sets server trace on or off. The name of the default trace file is soltrace.out. You can include one or more of the following tracing options
▪ sql: SQL messages
▪ sqlfetch: SQL fetch messages, prints info for each fetched row.
▪ sqlerrors: SQL errors
▪ syserrors: System errors
▪ est: SQL estimator information
▪ estplans: SQL execution plan
▪ rpc: Network communications
▪ rpcerrors: Network communication errors
▪ sync: Synchronization messages
▪ flow: Advanced Replication messages
▪ flowplans: Plans of SQL statements related to Advanced Replication
▪ rexec: Remote procedure call information
▪ batch: Background job and deferred procedure call information
▪ logreader: Logs the following information into the trace file soltrace.out.
– Logreader read started.
– Errors in Logreader cursor start. Total of 14 different error conditions are printed.
– Logreader read stopped.
– Abnormal read stop after certain system changes.
– High level information of number of returned log records and read progress.
Each information is tagged with the associated user id so that operations from different users can be separated.
▪ logreaderfetch: Rows fetched through Logreader interface.
▪ xa: Distributed transaction information.
▪ hac: High Availability Controller (HAC); trace information is output to hactrace.out in the HAC working directory. In general, solidDB support teams use the HAC trace facility for troubleshooting.
Using the HAC trace facility has minimal impact on performance.
To start tracing on HAC, you must run the command on a HAC connection. For example, connect to HAC with solsql using the port defined with the HAController.Listen parameter in the solidhac.ini configuration file.
▪ info level: SQL execution trace (level can be 0...8).
▪ func: Function execution information.
▪ proc: Stored procedure execution information.
▪ all: Both SQL messages and network communications messages are written to the trace file.
▪ active: Lists all active traces.
▪ lock: All lock requests including lock waits and timeouts.
▪ lockwait: Lock waits.
▪ locktimeout: Lock timeouts.
▪ noflush: Disables trace file flush after each write. Useful when tracing causes too much load on disk.
▪ latency: Latency of user operations. You can limit the output by using the limit option.
▪ sqllatency: Latency of SQL operations. You can limit the output by using the limit option.
▪ rpclatency: Latency of network operations. You can limit the output by using the limit option.
▪ limit usec: Latency limit in microseconds. Only operations that take more than usec microseconds are printed to trace file. Default is zero so all operations are printed.
▪ crep: Detailed information on replication, including execution of replication commands, detailed error messages, row-level replication progress, and connecting to the source database.