[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: PATCH fix 510970, 529551, 530541


On 11/23/2009 12:25 AM, Jerry Vonau wrote:
On Sun, 2009-11-22 at 15:48 -0600, Jerry Vonau wrote:
On Sun, 2009-11-22 at 21:27 +0100, Hans de Goede wrote:

On 11/22/2009 05:51 AM, Jerry Vonau wrote:


Ack, but your current split up is still not split up per purpose, I think
I understand what you are trying to achieve, so let me suggest a split:

1) A patch fixing the current copy code to use the fs with the most
      free space instead of the least

2) A patch to not only do the image transfer in case of a multiple disk
      cd / dvd install, but also in the case of a hdiso / net install, so as
      to free up memory used by install.img (which will be another patch).
      This patch should only do this when memory is tight! Doing this
      always is bad, as it is useless on systems with tons of memory, and
      could potentially even cause issues there with for example diskspace

3) A patch to actually achieve the freeing of memory 2)'s goal is by
      unlinking the install.img from /tmp

Think I got the flow you want for the patches now, I'm digging at the
free variable for the first part now. That will take me some time as I
don't have much to spare. What I have now is less that ideal, I would
like see what is being returned, and just use / for the image till the
above has being sorted out. At least for my testing, the rest of my
patches depend on that one being in place.

Oh, but that is easy, simply change:
          self._loopbackFile = "%s%s/rhinstall-install.img" % (anaconda.rootPath,
          self._loopbackFile = "%s%s/rhinstall-install.img" % (anaconda.rootPath,

Yea, I see that now, but I need to see what was being returned first ;-)

So that you use the last element of the list, instead of the first, so that
you get the biggest partition, your testing of this fix is much appreciated :)

That I can test quickly. I get back to you in a bit. (firing up vmm)

That works, I added some logging to return self._loopbackFile to my
updates.img, Then doing a vmm http install results in this logging:

Then a screen shot of df before/after

Thanks for testing, I've committed a slightly modified version of this patch
to the master branch, so atleast this will be fixed in F-13.



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]