From mboxrd@z Thu Jan  1 00:00:00 1970
From: Matthias Fischer <matthias.fischer@ipfire.org>
To: development@lists.ipfire.org
Subject: Core 128 Update - some glitches
Date: Fri, 15 Mar 2019 10:33:15 +0100
Message-ID: <e6ee0756-4ee7-9247-d441-c7f11b862f4f@ipfire.org>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============6776204623193898415=="
List-Id: <development.lists.ipfire.org>

--===============6776204623193898415==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

Hi,

yesterday I installed the update from Core 127 to Core 128 / 32bit on my
Duo Box with 4 GB RAM.

Pakfire showed *no* errors during the (second) try. But I noticed a few
"glitches":

1. First download attempt was aborted with:

***SNIP***
20:50:58	pakfire:	DOWNLOAD INFO: File received. Start checking signature...

20:50:59	pakfire:	DOWNLOAD ERROR: The downloaded file
(ipfire/pakfire2/2.21/paks/core-upgrade-2.2 1-128.ipfire) wasn't
verified by IPFire.org. Sorry - Exiting...

20:50:59	pakfire:	TIME INFO: Time Server 193.175.73.151 has -0.093144
sec offset to localtime.
***SNAP***

On the second try everything went smooth - absolutely no Pakfire or
installation errors - update went through.
Ok, it could have been me or my ISP. I didn't bother.

But after rebooting I saw that only the *SMP* kernel had been updated,
*PAE* is missing for some reason, I'm still on the old version. Can
anyone explain why?

Current contents of '/boot':

config-4.14.103-ipfire
config-4.14.86-ipfire-pae
grub
initramfs-4.14.103-ipfire.img
initramfs-4.14.86-ipfire-pae.img
lost+found
System.map-4.14.103-ipfire
System.map-4.14.86-ipfire-pae
vmlinuz-4.14.103-ipfire
vmlinuz-4.14.86-ipfire-pae

More details (german):
=> https://forum.ipfire.org/viewtopic.php?f=22&t=22463

So now I got a new SMP kernel (running) and an old PAE kernel...

Bug or feature or anything I can do about this!?

#####

2. I found that - not only with this but also with other updates - the
module 'nct6775' is installed on the Duo Box. I can't use this. This
module leads to sensor values being displayed in the hardware graphs -
with errors - which do *not exist* on the Duo Box.

Kernel says:
1 Time(s): nct6775: Found NCT6106D or compatible chip at 0x2e:0xa10

But there are no fans on this board...

Solution:

Stop 'collectd':
/etc/init.d/collectd stop

Delete directory:
'/var/log/rrd/collectd/localhost/sensors-nct6106-isa-0a10'

Unload 'nct6775':
/etc/modprobe.d/nct6775.conf

Delete 'nct6775' in '/etc/sysconfig/lm_sensors'

Start 'collectd' again:
/etc/init.d/collectd stop

I tried adding 'blacklist nct6775' to '/etc/modprobe.d/blacklist.conf'
but this had no effect.
Is there anything else I can do to prevent this module for being loaded?

#####

3. Today I noticed some kernel messages which I didn't see before after
an update:

***SNIP***
1 Time(s): <27>udevd[487]: RUN{builtin}: 'firmware' unknown
/lib/udev/rules.d/50-firmware.rules:3
1 Time(s): <27>udevd[487]: specified group 'input' unknown
1 Time(s): <27>udevd[487]: specified group 'kvm' unknown
1 Time(s): <27>udevd[487]: specified group 'render' unknown
1 Time(s): Kernel log daemon terminating.
1 Time(s): Kernel logging (proc) stopped.
***SNAP***

Besides these "problems", the machine is running smooth.
With ~3 GB RAM. ;-)

Can anyone confirm or has any hints?

Best,
Matthias

--===============6776204623193898415==--