Summary
archiving logic introduced in OA 11.03
Question
It’s a common scenario to see the following messages on status.scope file after logfile rollover:
**** /opt/perf/bin/scopeux : <timestamp> ****
ERROR: Unable to read from logfile '/var/opt/perf/datafiles/logxxx' - corrupted data. (PE221-24)
ERROR: Unable to read from logfile '/var/opt/perf/datafiles/logxxx' - corrupted data. (PE221-24)
And most of the OA users do not perform backups of the scope log* files, so whenever corruption happens all performance data is lost
Answer
With OA 11.03 we have introduced archiving logic:
1. Before performing logfile rollover, we archive logfile to:
a. logfile_timestamp.tar on all unix
b. logfile_timestamp on windows platform.2. If rollover is successful, we remove the archived file.3. If rollover results in corruption:
a. remove the corrupted file and start with a fresh logfile
b. user can use the archive file to extract data.
c. scope will still be running and starts logging data to the fresh logfile created.