Bug #23564
DAQInterface sometimes does not wait long enough for logfiles to appear
100%
Description
After starting the processes in an otsdaq-based system, I find that DAQInterface goes to the RECOVER state because it could not find the log file of one of the processes. A bit more debugging showed that the issue was related to the timing of the DAQInterface logfile check and the actual creation of the logfile; inserting a 5-second sleep was sufficient to ensure the file was found correctly.
It would probably be better for DAQInterface to retry within a certain timeout until all logfiles are found, and not just arbitrarily wait for some amount of time.
Associated revisions
History
#1 Updated by John Freeman 7 months ago
- % Done changed from 0 to 100
- Status changed from New to Resolved
With the single-commit-branch bugfix/23564_wait_for_logfiles, it's now the case that if DAQInterface determines a logfile's missing, it'll wait two seconds and check again, doing this for a maximum of five checks.
#2 Updated by Eric Flumerfelt 7 months ago
- Status changed from Resolved to Reviewed
- Co-Assignees Eric Flumerfelt added
- Experiment deleted (
-)
Code reviewed and tested. Merging into develop.
JCF: Issue #23564: when checking for logfiles, instead of a single check, check 5 times with 2 seconds between each check so they have plenty of time to appear