MinimServer Forum

Full Version: JAVA restart necessary to fix inactive Minimserver?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4
(10-06-2013 19:03)rb73 Wrote: [ -> ]
(10-06-2013 12:53)simoncn Wrote: [ -> ]I'll put a fix for this in the next release. Many thanks for your help with finding this problem!
You're welcome, glad I could help Shy
As requested in a PM from Simon, I herewith attach the minimserver-out-31833.log, the result of issuing the command:
kill -quit <pid>
pid being, in my particular case, 31833.
By the way, I had to change the .log extension to .txt. for it to be accepted by the upload mechanism.

Hope this helps!
(11-06-2013 07:11)hvaleton Wrote: [ -> ]As requested in a PM from Simon, I herewith attach the minimserver-out-31833.log, the result of issuing the command:
kill -quit <pid>
pid being, in my particular case, 31833.
By the way, I had to change the .log extension to .txt. for it to be accepted by the upload mechanism.

Hope this helps!

Thanks very much! This thread dump shows two "orphaned" instances of HTTPConnection$WriterThread. (The output from rb73 showed over 5000 of these.)

Please can you also attach the corresponding lsof output. This will enable me to cross-check and see whether there are leaked file descriptors corresponding to the "ophaned" threads.
Looks like it happened again sometime tonight, so I disabled MinimServer, copied the diagnostics archive, and enabled MinimServer again.

Diagnostics are in my dropbox again.
(11-06-2013 07:11)hvaleton Wrote: [ -> ]By the way, I had to change the .log extension to .txt. for it to be accepted by the upload mechanism.

I have updated the forum settings to accept .log attachments with a maximum size of 200 kB. Log files that are larger than this can be attached as .zip or .tar.gz files.
(11-06-2013 07:24)simoncn Wrote: [ -> ]
(11-06-2013 07:11)hvaleton Wrote: [ -> ]As requested in a PM from Simon, I herewith attach the minimserver-out-31833.log, the result of issuing the command:
kill -quit <pid>
pid being, in my particular case, 31833.
By the way, I had to change the .log extension to .txt. for it to be accepted by the upload mechanism.

Hope this helps!

Thanks very much! This thread dump shows two "orphaned" instances of HTTPConnection$WriterThread. (The output from rb73 showed over 5000 of these.)

Please can you also attach the corresponding lsof output. This will enable me to cross-check and see whether there are leaked file descriptors corresponding to the "ophaned" threads.
I've attached the corresponding lsof output. That is, I hope it's corresponding enough for you, because some 37 minutes elapsed between the kill and the lsof commands. If you want, I could do both again. But not until after my weekly round of sports....
(11-06-2013 07:29)rb73 Wrote: [ -> ]Looks like it happened again sometime tonight, so I disabled MinimServer, copied the diagnostics archive, and enabled MinimServer again.

Diagnostics are in my dropbox again.

The diagnostic files are different this time. The previous diagnostic files showed that MinimServer had crashed with an OutOfMemoryError. These diagnostic files don't show any errors, just that MinimServer was disabled while (apparently) running normally. This probably means there was a hang or deadlock, which might be related to the same resource leakage issue as the previous problem.

In the next release, I will enhance the diagnostic feature to include a Java thread dump. This will provide more information if MinimServer stops because of a hang or deadlock rather than an error. I will also exclude the .mcache files from the diagnostic archive, which will make the diagnostic archive considerably smaller.
(11-06-2013 07:39)hvaleton Wrote: [ -> ]I've attached the corresponding lsof output. That is, I hope it's corresponding enough for you, because some 37 minutes elapsed between the kill and the lsof commands. If you want, I could do both again. But not until after my weekly round of sports....

Thanks very much! This output appears to confirm the theory that the two issues are related.
Pages: 1 2 3 4
Reference URL's