Re: Problem source verified -Re: Jessie (8.0) slow to boot
On Thu 01 Sep 2016 at 13:11:22 (+0100), Darac Marjal wrote:
> On Thu, Sep 01, 2016 at 07:02:14AM -0500, Richard Owlett wrote:
> >On 8/31/2016 11:25 AM, Richard Owlett wrote:
> >>On 8/31/2016 10:44 AM, Felix Miata wrote:
> >>>Richard Owlett composed on 2016-08-31 09:58 (UTC-0500):
> >>>...
> >>>>Aug 31 09:13:32 deb8-2ndtry systemd[1]: Job
> >>>>dev-disk-by\x2duuid-0a344714\x2dae06\x2d43ed\x2daf89\x2d33ba51934630.de
> >>>>
> >>>>
> >>>>Aug 31 09:13:32 deb8-2ndtry systemd[1]: Timed out waiting for
> >>>>device dev-disk-by\x2duuid-0a344714\x2dae06\x2d43ed\x
> >>>>Aug 31 09:13:32 deb8-2ndtry systemd[1]: Dependency failed for
> >>>>/dev/disk/by-uuid/0a344714-ae06-43ed-af89-33ba5193463
> >>>>Aug 31 09:13:32 deb8-2ndtry systemd[1]: Dependency failed for
> >>>>Swap.
> >>>
> >>>Which was installed last, Squeeze, or Jessie? You can expect this
> >>>message trying to boot the earlier installed installation after
> >>>having done the later install. Unless you take affirmative action
> >>>to deny it, an existing swap partition used by the earlier
> >>>installation will be reformatted by the later installation.
> >>>Reformatting creates a new UUID, thus making the UUID referring
> >>>to it in the earlier installation's fstab invalid. The earlier
> >>>needs to have its fstab edited to use the correct swap partition
> >>>UUID, or volume label, or device name, if swap is actually
> >>>desired or needed.
> >>
> >>There may be some subtle problems still lurking somewhere which
> >>show up only for people like me doing many installs of *nearly*
> >>identical systems.
> >>
> >>Commenting out the line in /etc/fstab DID allow it to boot
> >>without warning messages.
> >>
> >>However,taking into account Darac's comments, does this now mean
> >>that the machines are now operating without a swap partition?
> >>
> >>IIRC the current instance of Jessie on the laptop was the
> >>chronologically the last OS installed so by the comments in this
> >>thread should not have had the problem. The laptop is my
> >>designated "Guinea pig" so I'll do a fresh install to see if
> >>problem persists.
> >
> >My procedure was:
> >1. Using the Live edition of Gparted, remove all partitions on
> >hard disk.
> >2. Using install DVD 1 of Debian 8.0.0
> > a. install Debian to /dev/sda1
> > b. create swap on /dev/sda2
> >3. Verify Debian boots without problems
> >4. Using the same DVD
> > a. install Debian to /dev/sda5
> > b. re-create swap on /dev/sda2
> >5. Attempt to boot both instances
> > a. Booting the install on /dev/sda1 generates the warning
> >message and use
> > of free shows swap does not exist.
> > b. Booting the install on /dev/sda5 generates no warning message
> >and use
> > of free shows swap does exist.
> >
> >I believe this justifies a bug report against the installer.
> >My expected behavior would be to check to see if a swap area
> >already exists before "creating" a swap partition. Especially
> >since replacing a swap partition can break a previously
> >functioning install in a multi-boot situation.
> >
> >Comments?
>
> I think that sounds sensible. A swap partition should be considered
> scratch space, and its contents only relevant to the
> currently-booted OS.
I disagree. If you want to share swap space, then you've got to
take appropriate action yourself. The easy way is to LABEL it in
/etc/fstab rather than UUID it, unless you like reading and typing
long strings of hex.
> The only fly in the ointment is hibernation. It's been a while since
> I played with hibernation, but I believe it stores the RAM image
> into Swap, so installing Debian while another OS is hibernated could
> even cause data loss. The solution there is A) to use a swap FILE
> for the hibernation image and B) for Debian to point-blank refuse to
> re-create a swap partition when it detects a hibernation image
> there.
I disagree. Is d-i expected to have a list of partition types, and
contents, that shouldn't be overwritten? Are all the flavours of fdisk
to test all the partitions on a disk before they allow you to
repartition it? If yes, then what about C> FDISK ? This just lulls
people into a false sense of security. Then Windows comes along...
Just as you can install 145 OSes on one PC, you can share swap space
among them, but it's then up to you to work out how to do this safely,
not just file a bug against an innocent installer doing what's asked
of it. The OP has enjoyed help from a sizeable thread entitled
"Problem with NON-STANDARD install" (his emphasis) which focussed
mainly on Grub. Perhaps we have to look forward to more problems
arising from the non-standard circumstances, but these will all
help with the OP's desired learning process.
Cheers,
David.
Reply to: