This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Summary: | [68cat][ruby][rake] Migrate:db peggs Dual CPU at 100% | ||
---|---|---|---|
Product: | ruby | Reporter: | esmithbss <esmithbss> |
Component: | Rake | Assignee: | Erno Mononen <emononen> |
Status: | RESOLVED INCOMPLETE | ||
Severity: | blocker | ||
Priority: | P3 | ||
Version: | 6.x | ||
Hardware: | PC | ||
OS: | Linux | ||
Issue Type: | DEFECT | Exception Reporter: | |
Attachments: | Profiler Trace from CPU 100% utilization. |
Description
esmithbss
2009-09-26 15:47:51 UTC
Created attachment 88398 [details]
Profiler Trace from CPU 100% utilization.
Multiple executions result in same performance problem so this appears to be reproducible. Migration file in question does two things: 1) drop a column from a table 2) change the name of a column within the same table. As in issue 173226 here it also looks like it's the external process that is pegging the CPUs - can you please do the same here, i.e. run the exact same command from the CLI and see how that performs? When I run the command from the IDE, it takes a horendous amount of time, but my CPU's aren't pegged. They are only running at about 50%. I don't know how the system is calling it's child processes, but NetBeans is behaving as if it isn't doing a spawn with a wait on the children, or a spawn with an async listener for the response, but instead is doing a spawn which redirects System.out and System.err to internal queues and has a very tight loop listening to the queues. It might be that the -client switch makes the difference here too; can you please check that? When you have the chance, can you please confirm whether the -client switch made any difference here. Thanks. |