[erlang-questions] etop flooded with "Erlang top got garbage" messages

Anton Lebedevich mabrek@REDACTED
Thu Mar 14 10:19:45 CET 2013

On 03/13/2013 06:26 PM, Anton Lebedevich wrote:
> On 03/13/2013 06:32 AM, Mike Cugini wrote:
>> Hey all,
>> I was recently put on an erlang project, and I've been trying to get
>> etop working to help profile some issues our app has been having (mostly
>> interested in Message Queue sizes).
>> Etop is started with this command:
>> erl -name etop@<ip> -hidden -s etop -s erlang halt \
>>                -output text -sort reductions -lines 50 \
>>                -node <node_name> -setcookie <cookie>
>> Whenever I start etop, I see a flood of "Erlang top got garbage" messages:
> -tracing off flag might help etop. Another possible source of that
> garbage is mnesia.

There is a correction to the last statement, thanks to Håkan Mattsson:

Mnesia doesn't produce that garbage by itself because it doesn't use
tracing. But it seems to trigger some bug in the way etop uses tracing.

It's reproducible with R14B04. Steps:

In one shell:
$ erl -boot start_sasl -name mns@REDACTED
(mns@REDACTED)1> mnesia:create_schema([node()]).

Then in another shell:
$ erl -name etop -hidden -s etop -output text -node mns@REDACTED

Observe normal etop output.

Then go back to the first shell and run:
(mns@REDACTED)2> mnesia:start().

Observe lots of these in the second shell:

Erlang top got garbage {trace_ts,<5149.95.0>,out,
Erlang top got garbage {trace_ts,<5149.91.0>,out,

Anton Lebedevich.

More information about the erlang-questions mailing list