I still stand behind my advice, but if doing this might solve some of the hang issues, it could be worth a shot. One option for testing the upgrade without affecting the site could be to use a backup snap shot to create clone, test the dist-upgrade, and then either swap out the old server with the new one, or then do the upgrade on the old one instead.

On Oct 16, 2013 10:57 PM, "Victor Boctor" <vboctor@gmail.com> wrote:
Paul has suggested that we upgrade server to a new version of the OS / kernel due to a kernel bug that is causing the hangs (is this proven?).  Last year (25-OCT-2012), I've asked whether we should upgrade to Ubuntu precise to John, Damien and Roland and the agreement was not to do it.

Paul, did you want to do something else on the server?

Are others aware of any other work that we need to get done on the server?

The other thing to discuss is the timing.  I would rather if we stay focused for the time being on getting 1.3.0a1 out of the door.  I don't want us distracted with environmental issues that are not critical causing us downtime, creating work, etc.

Here are the reasons:

** ROLAND **

I see no need to upgrade at the moment.
The system is running quite stable .

The version we are currently running (Ubuntu 10.04 LTS) is supported until April
"Precise" is the new LTS version with support until April 2017

We have 768MB memory available, during normal operation about 100-200MB free.
I didn't check, but most of the time newer versions have a larger memory
Maybe we would get worse performance caused by swapping.

I won't say "Never change a running system" but "Change running systems only if


As upgrading the server to 12.04 would likely require downtime, as
well as possibly cause a problem that would entail even more downtime
in cases where the dist-upgrade fails, I would agree with Roland in
saying no.  If there is a significant benefit to the newer release,
and we do decide to run a dist-upgrade, I would just like to remind us
that putting the site into a read-only mode beforehand would be a good
precaution, as would running a manual Linode backup.  If something
does fail during the upgrade, reverting to the latest Linode snapshot
would be the best bet.


Silence :) but he agreed based on follow up discussion.



October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
mantisbt-dev mailing list