Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tracecompass-dev] Question about VM tracing with LTTng and Virtual Machine Analysis inside TraceCompass

Hi Emilio and Dario,


Thanks for the interest in the virtual machine analysis feature. This documentation is something I have not revisited in a few years, so it is indeed outdated.


As mentioned, for the trace synchronization, the newer lttng versions have support for it out of the box now. The doc should only mention that.


As for the virtual machine analysis itself, I worked on an approach more straightforward, if you are using Qemu/KVM. Here's a presentation from almost 3 years ago on the topic, linking to more recent (but still outdated) lttng-tools/ust/modules branches [1]. This approach does not require trace synchronization at all.


The Trace Compass documentation was never updated to link to these, as I was hoping they would make it in upstream lttng first and was waiting for that. But that never happened, due to lack of interest by the community.


If you are interested, I can try to see if I can quickly rebase these branches on the latest lttng 2.12, so you can try it out. You could also contact the lttng community to see if they have anything in their roadmap regarding this feature.


[1] https://amdls.dorsal.polymtl.ca/system/files/ProgressMeetingMay2018_gbastien.pdf



Best regards,

Geneviève



On 2021-02-16 11:38 a.m., tracecompass developer discussions wrote:
Hello,

We are trying to trace a VM with LTTng and then make the Virtual Machine Analysis inside TraceCompass. Following the documentation instructions ( https://archive.eclipse.org/tracecompass/doc/stable/org.eclipse.tracecompass.doc.user/Trace-synchronization.html#Trace_synchronization ) for synchronization of the traces, we are encountering some errors:

- In the section "LTTng-module network tracepoint with complete data", it seems that the git repository mentioned there does not exist any longer:

Cloning into 'lttng-modules'...
fatal: unable to connect to git.dorsal.polymtl.ca:
git.dorsal.polymtl.ca[0: 132.207.72.58]: errno=Connection refused

- In the section "LTTng-modules addons kernel module with dynamic tracepoints", when compiling the modules, we get the following errors:

cc1: some warnings being treated as errors
make[4]: *** [/usr/src/linux-5.10.12-1/scripts/Makefile.build:280:
/home/emilio/old_lttng/lttng-modules/probes/lttng-probe-sched.o] Error
1
make[3]: *** [/usr/src/linux-5.10.12-1/scripts/Makefile.build:497:
/home/emilio/old_lttng/lttng-modules/probes] Error 2
make[2]: *** [/usr/src/linux-5.10.12-1/Makefile:1819:
/home/emilio/old_lttng/lttng-modules] Error 2
make[1]: *** [../../../linux-5.10.12-1/Makefile:185: __sub-make] Error
2
make[1]: Leaving directory '/usr/src/linux-5.10.12-1-
obj/x86_64/default'
make: *** [Makefile:110: modules] Error 2

(full output available here: https://pastebin.com/EP1ZDiv4)

Finally, when trying to build the Qemu/KVM extra tracepoints ( https://archive.eclipse.org/tracecompass.incubator/doc/org.eclipse.tracecompass.incubator.virtual.machine.analysis.doc.user/Virtual-Machine-Analysis-User-Documentation.html ), we also face build errors:

cc1: some warnings being treated as errors
make[4]: *** [/usr/src/linux-5.10.12-1/scripts/Makefile.build:280:
/home/emilio/old_lttng/lttng-modules/probes/lttng-probe-sched.o] Error
1
make[3]: *** [/usr/src/linux-5.10.12-1/scripts/Makefile.build:497:
/home/emilio/old_lttng/lttng-modules/probes] Error 2
make[2]: *** [/usr/src/linux-5.10.12-1/Makefile:1819:
/home/emilio/old_lttng/lttng-modules] Error 2
make[1]: *** [../../../linux-5.10.12-1/Makefile:185: __sub-make] Error
2
make[1]: Leaving directory '/usr/src/linux-5.10.12-1-
obj/x86_64/default'
make: *** [Makefile:111: modules] Error 2

(full output available here: https://pastebin.com/dcjCNf43)

Is there a more updated version of those repositories that we can use, or is the process for doing a VM analysis different from what is described in those two sections of TraceCompass documentation?

Thanks and Best Regards,
Emilio Bruno, Dario Faggioli

_______________________________________________
tracecompass-dev mailing list
tracecompass-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/tracecompass-dev

Back to the top