Process launch/terminate logging and Firefox (extension?) issue

Started by muse2u, September 18, 2016, 03:16:33 PM

Previous topic - Next topic

muse2u

I noticed that this issue was reported in 2014 and two years later is still a problem. There is some suggestion that there will be a fix in v9. Let me add my voice to a request for a solution sooner than later. ADMIN EDIT: NO, it was not. He was speaking of a different issue.

My primary use of Process Lasso for the last few years is as a tool to monitor the processes on my Windows 10 computer and to perform troubleshooting. The information provided by Process Lasso is far above anything else that I have seen in other task/process monitors, and has been the tool of choice in resolving many performance issues. The value in its use as a performance monitor tool is proportional to the confidence in its reported information. If the information is misleading or inaccurate, or the monitor/logging function exacerbates the problem, the program is useless as a trouble shooting tool.
I have read the forum entry several times, and since I am not as well versed in the subject as your moderators, I have only a limited understanding.

My experience with this issue is as follows:
In the last year or so, I have seen logs of the events described above with respect to Firefox. Process Lasso has log settings for process launch and terminate turned on. A launch entry is logged at Firefox start, and Firefox executes normally for some time. Something then triggers a Firefox/Process Lasso issue, that reflects, what appears as a loop in the logging process, an endless logging of process launch (with the same PID) and process terminates (with the same PID but different PID as the launch). The counter of Firefox uptime increments, suggesting that the process is still running.

In the first incarnation of this problem, the looping log entries were accompanied by a significant performance problem that was not identified by Process Lasso (i.e. according to Process Lasso responsiveness was 100% when it clearly wasn't). Terminating any Firefox processes using Process Lasso, did not clear the performance issue nor the log loop. With Firefox processes no longer in the Process Lasso task list, it was necessary to bring up the Windows Task Manager, which showed that Firefox was still executing. Terminating Firefox with the Windows Task Manager, cleared both the performance issue and the Process Lasso log loop.

In the second incarnation of this problem, a few months ago something happened to change the experience of the problem (as commented above, maybe a Windows 10 update). The Process Lasso log looping was no longer accompanied by a performance issue and, after terminating Firefox processes in Process Lasso, the Windows Task Manager failed to show any running Firefox tasks. So, as was noted above, maybe the running Firefox task was 'hidden' and could no longer be manually terminated, or the task had ended and Process Lasso logging was hung in the log loop.

In both scenarios, stopping and starting Process Lasso and the governor, had no effect on clearing the loop, only a reboot of the computer cleared the problem.

It should also be noted that while the Firefox task is up and running (before any process termination is performed manually) and the loop is being registered by Process Lasso, Firefox is useable (somewhat sluggish in scenario 1) and therefore I don't understand the conclusion that the loop is caused by a crashing process, ie. Firefox crashing.

The suggestion above that we turn off logging of process launch/terminates just hides the problem and is not a solution. The option is not selective, it prevents logging of all process launch/terminates, information that is valuable in performance identification troubleshooting. If the log loop is a figment of Process Lasso imagination or is actually caused by the running of Process Lasso, then we have a Process Lasso confidence issue. If it is reporting an actual Windows process failure, it is doing the job that I paid for.

Does anybody have any idea where in the Bitsum scheme of things, the status of the solution is? Two years is a long time between identification of a problem and a solution in a V9 version of the program!

   

Jeremy Collake

New consolidated response for email subscribers:

Quote from: muse2u on September 18, 2016, 03:16:33 PM
It should also be noted that while the Firefox task is up and running (before any process termination is performed manually) and the loop is being registered by Process Lasso, Firefox is useable (somewhat sluggish in scenario 1) and therefore I don't understand the conclusion that the loop is caused by a crashing process, ie. Firefox crashing.

That is because you are reporting an entirely different issue that looks the same (while accusing me of not acting for 2 years.. if only you knew how hard I work to keep Bitsum here).

We are going to keep an eye on Firefox and see if we can reproduce this, but this is the first report we've had, and haven't had a 'crashing software loop' report in many months, so it's unclear if that issue still exists.

As for confidence in Process Lasso, you should be supremely confident in it. BUT people are running extremely complex software ecosystems in their Windows installs, so it is hard to achieve 100% perfection in all scenarios just due to this software ecosystem complexity (e.g. the security software interoperability issue fix this week).

It would appear from your description that you have a something else going wrong on your PC. We have testers running Firefox constantly.

1. The PIDs you report as changing (on new launches).
2. Restarting Lasso doesn't fix the issue (clue that it's not Lasso)
3. I think you have a Firefox Extension that may be crashing ...
4. In any event, it is definitely *not* the 2 year old issue reported here, and I am still a little offended by the implication that I would rest of my laurels for 2 years if a critical bug did exist, and was known. I just work so hard, anyway... It's a late Sunday night, and I'm splitting this thread as it's a DIFFERENT issue.
Software Engineer. Bitsum LLC.

BenYeeHua

QuoteWe are going to keep an eye on Firefox and see if we can reproduce this, but this is the first report we've had, and haven't had a 'crashing software loop' report in many months, so it's unclear if that issue still exists.
Not sure, I don't keep an eyes on this crashing software loop, as I am bit busy now...
If it happen again, I will checking more information for it, but normally it is useless, as it just went hidden for other Task Manager...

Jeremy Collake

As I described above, it sounds A LOT like either an extension gone wrong OR an extension legitimately launching background instances of Firefox.

In either scenario, Process Lasso is reporting what it is seeing and it is not related to the 'crashing process' logging issue.

I do not think it's a bug (unlike the temporary crashing process issue), but it could be handled better.

@BenYeeHua: Please continue to report on THAT bug in the other thread, as you did earlier.
Software Engineer. Bitsum LLC.