X2Go Bug report logs - #465
possible memory leak; throughput problem

Reply or subscribe to this bug.

Toggle useless messages

Message #5 received at submit@bugs.x2go.org (full text, mbox, reply):

From: Brian Lim <brian.lim.ca@gmail.com>

To: submit@bugs.x2go.org

Subject: possible memory leak; throughput problem

Date: Fri, 28 Mar 2014 01:49:30 -0400

[Message part 1 (text/plain, inline)]

Package: x2goserver
Version: 4.0.1.13

I am running a Java program that performs a timer initiated operation once
per minute throughout the work day. The elapsed time of this operation is
usually about 2 milliseconds. I am viewing it remotely with either X2go or
tightvnc. When I use tightvnc, I experience no problem. When I use X2go,
there is no problem for the first 200 to 300 minutes, but at some point in
time it will start to experience long elapsed times. Elapsed times increase
gradually from about 2 milliseconds, to a few seconds, to even a few
minutes. The maximum elapsed time varies. Then the elapsed time will drop
back to 2 milliseconds and the pattern repeats.

I have been using tightvnc since 2012 and only in March of 2014 did I
switch to X2go. It is then that I noticed this problem.

This could be a memory leak problem.

The server is running Ubuntu 12.04 LTS and the command x2goversion reports
the following.

x2goagent: 3.5.0.22
x2goserver: 4.0.1.13
x2goserver-extensions: 4.0.1.13

The client is also Ubuntu 12.04 LTS and the version of x2goclient is
implied by the version of ubuntu.

Attached is a LibreOffice document containing elapsed time plots.

Brian Lim

[Message part 2 (text/html, inline)]

[x2go memory bug report 2014 march.odt (application/vnd.oasis.opendocument.text, attachment)]

Message #10 received at 465@bugs.x2go.org (full text, mbox, reply):

From: Brian Lim <brian.lim.ca@gmail.com>

To: 465@bugs.x2go.org

Subject: additional information

Date: Fri, 28 Mar 2014 02:19:27 -0400

[Message part 1 (text/plain, inline)]

I am running with Java's -Xincgc option which enables incremental garbage
collection. It is my understanding that incremental GC sacrifices
efficiency in order to gain consistency. In other words, there is more work
to do but since it is being done frequently, there should not be large
elapsed times imposed by having to do a very large GC operation. Since my
application has time to spare (at least when I use tightvnc), this is my
preference.

Since discovering this problem, I have increased the memory of the server
from 1.5 GB to 2 GB and I have increased the memory allowance for the Java
applications but the problem persists.  I may report back on performance on
a 4 GB server next week.

Brian Lim

[Message part 2 (text/html, inline)]

Send a report that this bug log contains spam.