Problems with Apper / automatic upgrading of my Debian 7.11 system

[rhkr...@gmail.com]

Background:

Since I upgraded to Debian 7.11 (wheezy) maybe 6 months to a year ago (maybe
longer), I've made it a habit to upgrade any software for which Apper notifies
me that an upgrade is available.

Something like two weeks ago, I upgraded the Linux kernel when notified.

Before the upgrade completed, I got the following popup:

<quote>
Package failed to install - Apper

One of the selected packages failed to install correctly.
More information is available in the detailed report.
</quote>

And, upon clicking on Details, I got the following:

<quote>
subprocess installed post-installation script returned error exit status 1
</quote>

But, the upgrade seemed to complete successfully, eventually giving me the
message (~"Your system is up to date, verified <n> hours and <m> minutes ago").

Aside: I did not reboot after that upgrade nor any time since then.

After that, the next several upgrades proceeded in the same fashion--that is,
I saw the warning that one of the selected packages failed to install
properly, but it then seemed to complete and said my system was up to date.

Since then (maybe the last 2 or so upgrades, in the last 2 or 3 days) the
upgrades do not complete successfully--the apper screen just "hangs"--I do not
get the message that the upgrade completed successfully, even if I let the
apper screen sit there for half an hour.

Today I got a notification to update default-jre, default-jre-headless, and
java common and:

* it gave me a list of other software that had to be updated: libsctp1,
openjdk-7-jre, and openjdk-7-jre-headless, and I accepted that (clicked
Continue)
* it asked me for authentication, which I gave it
* it downloaded the packages
* it started preparing and installing the packages

The popup "Package Failed to install - Apper" appeared-- this time I noticed
that it appeared during an attempt to install "linux-image-3.2.0-4-amd64 Linux
3.2 for 64-bit PCs".

I don't know (didn't notice) if during the other problem installs it attempted
to install the linux image.

(Aside: I assume that a new linux-image is only activated after a reboot?
OTOH, when I run uname -a I get: "Linux s19 3.2.0-4-amd64 #1 SMP Debian
3.2.78-1 x86_64 GNU/Linux" which seems to indicate I am using the latest Linux
kernel (??).)

Anyway, it seemed that today's update did complete successfully--after today's
upgrade I did get the message:

<quote>
Your system is up to date
Verified 3 hours and 26 minutes ago
</quote>

So:
1. Should I be worried?
2. Should I reboot?
3. Was there something wrong with that kernel upgrade that I should expect
to be fixed, and wait to install that fix before rebooting?
4. Can anyone explain what is going on?
5. Any other advice?

Thanks!
Randy Kramer

[rhkr...@gmail.com]

Thanks for the reply, but I want to clarify a little more (oh, and the other
questions still stand...):

* I assume that I only have to reboot if I / when I'm ready to have that
new kernel be activated--otherwise the system will continue to use the old
kernel--presumably with no problems.

* Or, do you think my problems with upgrading / Apper are because I didn't
activate the new kernel?

* But, that wouldn't explain the problem with they first problematic
upgrade, which was the kernel...

Thanks for any clarification you can add...

On Monday, June 27, 2016 10:27:24 AM somebody who didn't want to reply on list
wrote:
> If the kernel was updated, you MUST reboot. And that's the only case
> where it is mandatory.

[Johann Klammer]

Don't reboot until you know that the bootloader got installed correctly.
Ideally have a boot CD ready in case it goes wrong.

[rhkr...@gmail.com]

That sounds like good advice, thanks!

But, has anyone else had similar problems recently with "automatic" upgrades
in apper (see my original post), or can anyone explain (or theorize) on what
is going on?

[rhkr...@gmail.com]

Ok, I now believe that my problem is that the last Linux image (kernel) update
did not install correctly / completely. That image was "linux-image-3.2.0-4-

amd64 Linux 3.2 for 64-bit PCs"

I believe that what has been happening since then is that, each time I've
installed something else (either recommended by apper to keep the system up-
to-date or as a new program I wanted) both the intended program and the Linux
image attempted to install, and, typically, the intended program was installed
or updated successfully, but the Linux image was not.

I can't absolutely confirm that for every case before a few days ago, but, in
the updates or installs I've done since then, I've noticed that the popup
message that says there was a problem occurs while apper is attempting to
install the new kernel. (To clarify--since that attempt about two weeks ago
to install the kernel, every subsequent attempt to install anything has caused
that message to popup, but, before a few days ago, I hadn't noticed that apper
was re-attempting to install the Linux image).

Aside: apper has installed several linux images before this one, and I never
had this problem before, and typically did not reboot soon after the update--
instead, I waited until there was some other reason to reboot.

So, now what do I do?

Per the warning from Johann Klemmer, I want to be careful about rebooting--I
don't want to attempt a reboot that might fail.

Is there any log of the apper install process that would give me more
information than the popup messages:

<quote>
Package failed to install - Apper

One of the selected packages failed to install correctly.
More information is available in the detailed report.
</quote>

And, upon clicking on Details, I get the following:

<quote>
subprocess installed post-installation script returned error exit status 1
</quote>

Other suggestions?

Thanks,
Randy Kramer

Aside: I suppose this might be considered somewhat rude, but I wrote the above
with a headache, my glasses off, and only a cursory attempt at proofreading--I
apologize!

On Monday, June 27, 2016 01:48:41 PM Johann Klammer wrote:

[Selim T. Erdoğan]

On Tue, Jun 28, 2016 at 09:32:43AM -0400, rhkr...@gmail.com wrote:
> Ok, I now believe that my problem is that the last Linux image (kernel) update
> did not install correctly / completely. That image was "linux-image-3.2.0-4-
> amd64 Linux 3.2 for 64-bit PCs"
>
> I believe that what has been happening since then is that, each time I've
> installed something else (either recommended by apper to keep the system up-
> to-date or as a new program I wanted) both the intended program and the Linux
> image attempted to install, and, typically, the intended program was installed
> or updated successfully, but the Linux image was not.
>
> I can't absolutely confirm that for every case before a few days ago, but, in
> the updates or installs I've done since then, I've noticed that the popup
> message that says there was a problem occurs while apper is attempting to
> install the new kernel. (To clarify--since that attempt about two weeks ago
> to install the kernel, every subsequent attempt to install anything has caused
> that message to popup, but, before a few days ago, I hadn't noticed that apper
> was re-attempting to install the Linux image).
>
> Aside: apper has installed several linux images before this one, and I never
> had this problem before, and typically did not reboot soon after the update--
> instead, I waited until there was some other reason to reboot.
>
> So, now what do I do?

I sometimes have problems with /boot getting full. If you've installed
several linux images and they're all taking up room in /boot, you might
have to get rid of an old one (which you won't be using anymore).

You can check if /boot is getting full with the "df -h" command.

To get rid of an old image cleanly, just purge the associated package.

[Gene Heskett]

Another case where one can get bit is if the vmlinuz file is split by the
allocation of a 2nd inode to it, and this inode is beyond the ability of
the bios to reach. My own /boot is a directory in /, and without a
separate /boot partition, this could bite me even though a du -h /boot
says its only using 97Mb.

Our tools for mapping that and showing us meaningful data do not seem to
exist, or I have not been made aware of them.

Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

[David Wright]

That might prevent you from booting a given kernel, but it shouldn't
prevent you installing anything successfully.

> Our tools for mapping that and showing us meaningful data do not seem to
> exist, or I have not been made aware of them.

Is hdparm --fibmap what you are missing?

Cheers,
David.

[Gene Heskett]

Looks good David, except it is a file at a time, not liking a /boot/*
invocation. Works well on individual files.

Thanks for the clue, now I should know how to check it the next time I
think its biting me.

[rhkr...@gmail.com]

Thanks to all who responded! The problem is resolved. The basic problem was
lack of space on /var/tmp. I think the biggest takeaway (for me) was the
suggestion (by Steve) to run apt-get upgrade which gave me a much more
informative error message (than apper).

Steve <dl...@bluewin.ch> wrote to me offlist and suggested that I run apt-get
update and then apt-get upgrade. I did that, and apt-get upgrade told me
there was one package not installed and offered to install it--I said yes and
got a much more informative message than the one(s) I got from apper.

The install failed do to a lack of space on /var/tmp. I found that that the
kde http cache was using a lot of space (over half of the 500 mb partition--I
have it on a separate partition for, for me, historicial reasons)
(/var/tmp/kdecache-rhk/http/).

I removed all the files from that directory (rm -rf /var/tmp/kdecache-
rhk/http/*), retried apt-get upgrade, and the linux-image was installed
successfully. (I still haven't rebooted, and don't plan to until there is
some good reason to do so.)

So far I haven't noticed any problem that might be due to clearing the kde
http cache, nor do I expect any, except perhaps some delay in re-viewing
existing open tabs (which I haven't noticed so far).

(Clarification for Steve (note ot others: in responding to Steve I mentioned
that I thought the problem was the growth of .xsession-errors, and somehow, I
thought I had found an .xsession-errors on /var/tmp--I must have done
something wrong (like typed in the wrong directory). In any event, .xsession-
errors had grown to some outrageous size, and I've now put commands in my
crontab to truncate that file once a minute. Unfortunately, the space hasn't
been freed up yet, but it is irrelevant to the problem of installing the linux
image. Of course, rebooting is one of the ways (iiuc) to free up that space,
but I think, sometime in the past, I found another way to do that, and think
(hope) I have notes about how I did that--I will look for those before
resorting to rebooting.)

Thanks again to all who responded! There were other useful suggestions in
this (and a sister) thread that I intend to add to my notes for potential
future use.

regards,
Randy Kramer

PS: Apologies--writing this with a headache and without careful proofreading.

PPS: Someone thought that one of my statements was unclear--iirc, it had to do
with previous Linux images that I had installed without problems--iirc, I was
trying to say that those Linux images had, according to apper, installed
properly, but I had never rebooted. (uname says my system is using version
3.2.78-1of the Linux kernel, and this upgrade was for 3.2.81-1--so I'm
guessing that versions .79 and .80 exist somewhere on my system (or may have
existed), but were never actually activated because I haven't rebooted.


On Wednesday, June 29, 2016 05:36:28 PM Gene Heskett wrote: