[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tracecompass-dev] [lttng-dev] Question related to LTTng

Hi Ravindra,

On 2019-03-27 2:49 p.m., Ravindra Kumar Meena wrote:

> So you have a user space application that generates a trace. It is virtualized, and you want to open it in trace compass to analyze it.


Yes. That's what I want to do but it has to in real-time manner.

Trace Compass does not support live traces. It is made for post-mortem analyses, so it works only on complete trace. We briefly supported live traces a few years back, but that made the code much more complicated, so this support was dropped. TraceCompass is not made for trace monitoring!

> I would like to know if there is a way to transfer CTF to Trace Compass in a real-time manner using TCP/UDP.

>Would scp work? Just asking.


I am supposed to transfer it through only TCP/UDP.

LTTng does support relaying the data over the network (see https://lttng.org/docs/v2.10/#doc-sending-trace-data-over-the-network). But Trace Compass does not open traces that are not terminated. The 2.11 version (not yet released) and master also support session rotation, so you can have traces in chunks of say 1 minutes and whenever a chunk is closed, it is ready to be opened by Trace Compass. This is as close to live trace analysis as you can get.


>How can I convert CTF Trace Data to LTTng? Since TraceCompass already understands LTTng Trace Data.

>The CTF trace should be openable in Trace Compass. You won't have as many pretty graphs and whatnot, but you can get some basic analyses done with searching and filters. If you want some more advanced analyses, you can code an XML analysis, or use any language you want and parse it to make a LAMI report. Finally you can make your own analysis (and even commit it to the incubator! ð ð ).

  >The information that I want to analyze and display information includes CPU usage, IRQ analysis(IRQ Statistics, IRQ Table, IRQ vs Count, IRQ vs Time), Linux Kernel(Control Flow, Resources)

With an LTTng kernel trace, you can get all that. Add a UST trace to it and you can correlate both traces together.