Re: System cannot come back from suspend (Fedora 9)

Below, the "at" symbol (Shift-2) has been changed to "(at)" so that googlegroups
won't render those parts of the post unreadable.

Mark Roberts <markrobt+usenet(at)> wrote:
From: Mark Roberts <markrobt+usenet(at)>
Newsgroups: comp.os.linux.misc
Subject: Re: System cannot come back from suspend (Fedora 9)
Date: 12 Oct 2008 00:33:08 GMT
Organization: 1.94 meters
Lines: 47
Message-ID: <6lcuu4Fbi7utU1(at)>
References: <6l80tqFaeaupU1(at)> <UXLHk.8607$jZ3.7453(at)newsfe03.iad> <6lae72Fbe6plU2(at)> <2fXHk.2776$0e6.165(at)newsfe09.iad>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: t5s5EM+DvQ8BulkYllI2TAz5EziY3CagHbB9bXq+XHeGEYsuOt
Cancel-Lock: sha1:Vq/oe7wcU5l+S0eliXyzsunqzAM=
User-Agent: slrn/0.9.9p1 (Linux)

Stefan Patric <not(at)> had written:
| Since you did "turn off" your machine, I had to ask. If your system had
| booted normally instead of hanging, it would have probably done a fsck
| anyway. But it never hurts to be thorough.

As it turns out, the system, when booted with the
kernel, continues to be flaky. Sometimes it can boot up and then
just freezes in the middle of something; other times, it doesn't
even complete the boot sequence.

I have now run fsck on all disks. They consistently come up clean
other than a "Superblock last mount time is in the future" message.
I've done mkswap on the swap partition.

| > If there were dependencies, rpm -e wouldn't have let me uninstall the
| > kernel, right? (At least not without --force, which I didn't use.)
| > Normally, I do use yum. But there are a few instances for which it is
| > still good to know and use rpm.
| That's not my understanding of how rpm works: you rpm -e some package
| and it removes everything that was in the original installed package.

As it turns out on that point, you were correct and I was wrong. (I
did a little experimenting on a package that I don't care about.)
So it's possible that I nailed something in the kernel, though I'd
be surprised given that the kernel always installs as a new
install. Trying it again with
'yum erase kernel-' finds no dependencies. But
I have removed it and re-installed it on the outside chance
that it will make a difference.

Interesting warning message with those steps:
"resolveDevice: device spec expected". It did not prevent the
installation from completing.

The one new clue that I have is in /var/log/pm-suspend.log -- where
the original suspend operation is shown as "Save video state
failed". That may be the root cause of the problem, but I think it's
gone way beyond that point now considering all the other flakiness that
I'm seeing when using this kernel.

Mark Roberts - Oakland, CA - NO HTML MAIL
Permission to archive this article in any form is hereby explicitly denied.
If you quote, please quote only relevant passages and not the whole article.

This one too, is for the record.