Linux

As scheduled to end of October 2024: Announcing Fedora 41 from Red Hat.

Such distro has mechanism to do in-place upgrade: Upgrading Fedora Linux Using DNF System Plugin. It is based on DNF System Upgrade.

Priciple is to run dnf system-upgrade download --releasever=41, tap Yes couple times and then dnf system-upgrade reboot. It works (mostly) and I have used such in-place upgrade many times on a VM running in Hetzner.

If you haven't read between the lines yet, let's state the obvious: I'm posting this in a scenario where everyhing didn't go as planned.

Imagine a virtual machine running in a data center far far away. There is interaction via SSH or if needed, a browser-based console can be used for dire needs. A failed update was indeed such.

41 Upgrade Begins - And Fails

Simultaneously with dnf system-upgrade reboot, I start sending ICMP echo requests to my VM to see the point in time when it begins pinging me back. This is a clear indication of upgrade being finished. I waited roughly 20 minutes without a response. Such a long time is an obvious indicator of a problem. Subsequently I logged in Hetzner's portal to pop open a console. Console showed me an upgraded system in the middle of a reboot-cycle. Stuck doing nothing.

That being unexpected, I chose to go for a Ctrl-Alt-Del. My wish came trough, a request to reboot nicely recycled the stuck system and a login-prompt greeted me on the console. Still, ping didn't. On the console, the only single keyboard layout made available is hard-coded ANSI US. On hardware, all my keyboards have layout ISO Finnish. That makes those elusive often used characters like dash (-), slash (/), pipe (|) or colon (:) be in very very different places slowing the entire process.

On the Console - Missing Package

Poking around the system on console indicated an upgraded VM. Everything else on the system was nice & peachy besides networking. There was no IP-addresses assigned. Actually entire NetworkManager was missing from the system. It did not exist. At all! Also every single bit of configuration at /etc/NetworkManager/ was absent.

Transferrring the muich-needed RPM-package NetworkManager-1.50.0-1.fc41 by eyeballing a rather dumb virtual console is fruitless. A quick analysis of the thing ended with a harsh one: it doesn't support any sensible means of transmitting files. Receiving any sent data with copy/paste or any other low-level means was absent. Something else was needed.

The Fix - Scraping

I opted to fix the network by hand. ip-command was installed in the system and it worked perfectly. That's all I needed. Or, almost all.

In my infinite wisdom, I didn't have any of the IP-details at hand. I reasoned to myself the system upgrade having worked flawlessly multiple times before this. I didn't NEED to save IPv4 or IPv6 -addresses, their routing setup or DNS-resolvers. I knew from tinkering with these boxes that on x86-64 architecture Hetzner VMs  all those details are static, manually set to Hetzner-assigned values. Their modern setup on Arm v8 does utilize DHCP for IPv4. My box was on a traditional rack and I couldn't count on automation to assist on this one.

Scraping all the bits and pieces of information was surprisingly easy. My own DNS-records helped. After the fact, I realized a shortcoming, if I would have looked at the bottom of the web-console, those IP-addresses would have been shown there. At the time I didn't. Routing defaults can be found from documentation such as Static IP configuration.

Now I knew what to set for the values.

The Fix - Manual Labor

Now the "fun" begun. I need to setup IPv4 address complete with routing to restore functionality of dnf-command. This would make it possible to install NetworkManager to get nmcli-command back.

Sequence is as follows:

ip addr add 192.0.2.1/32 dev eth0
ip route add 172.31.1.1 dev eth0 src 192.0.2.1
ip route add default via 173.31.1.1 src 192.0.2.1

Btw. see RFC5737 for IPv4-addresses and RFC3849 for IPv6-addresses reserved for documentation. I'm not going to post my box's coordinates here.

Fedora DNS-setup is via systemd-resolved, checking file /etc/systemd/resolved.conf. File had survived the update intact. It still had the content of:

DNS=185.12.64.1 185.12.64.2 2a01:4ff:ff00::add:1

A perfect & valid solution.

The Fix - Managing Network

Ta-daaa! Ping worked. dnf worked! Everything worked! The joy!

At this point running dnf install NetworkManager wasn't much. Trying to figure out what was wrong proved to be a headache.

On initial glance nmcli conn show:

NAME  UUID                                  TYPE      DEVICE
eth0  12345678-1234-1234-1234-123456789abc  ethernet  --

What!? Why isn't my eth0-connection associated with a device! No amount of attempts, tinkering, cursing nor yelling helped. I could not associate a device with the connection. My only option was to drop the hammer on the thing: nmcli conn del eth0

Now my eth0 didn't work as it didn't exist. A delete made sure of it. Next, getting it back:

nmcli conn add type ethernet ifname eth0 con-name eth0 ipv4.method manual ipv4.addr 192.0.2.1
nmcli conn modify eth0 ipv4.gateway 172.31.1.1
nmcli conn modify eth0 ipv6.addr 2001:db8::1/64
nmcli conn modify eth0 ipv6.gateway fe80::1

Final twist was to make those changes effective: nmcli device reapply eth0

IPv6 begun operating, IPv4 was unchanged. Everything was where it should have been after the upgrade.
That was it for NetworkManager, moving on.

Outcome

The only true measure of a possible success is a system reboot. If my tinkering survived a normal cycle, then all was good. Nothing special to report on that. Everything stuck and survived a rinse-cycle. File eth0.nmconnection stored NetworkManager configs as expected.

Why this thing exploded remains unknown. Missing any critical pieces of a system is always a disaster. Surviving this one with very little damage was lucky. I may need to increase my odds and re-image the VM. My guess is, this box survives only so-and-so many upgrades. I may have used all of the lives it has.

My previous post was about 10+ year old laptops. At that time Broadcom ruled the chipset Wifi chipset market.

Since those days balance has shifted. Today, most common chipset for Wifi is Realtek. I also have couple Mediatek Wi-Fi chips working perfectly with Linux.  To repeat what I said previously: These guys have their Linux support via open-source drivers. Broadcom doesn't. Hm. I dunno, maybe that's what made them decline and the other guy thrive? Most certainly, I wish it was their open-source support. 😁

So, my old laptop lost wireless connecitivy and I needed SOMETHING, to get the thing going to The Internet. I happened to have a brand new USB-stick on a test drive. As there aren't many Linux-supported chipsets, most USB-sticks won't work with 802.11ax / Wi-Fi 6 and you have to settle for slower speeds combined with less security. This product is supposed to beat compeition on that.

Spoiler: I doesn't! (yet)

The product is BrosTrend AX1. The speed rating is AX1800 and it's supposed to be WiFi 6 Linux Compatible WiFi Adapter:

lsusb information:

Bus 001 Device 013: ID 0bda:b832 Realtek Semiconductor Corp. 802.11ac WLAN Adapter
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               2.00
  bDeviceClass            0 [unknown]
  bDeviceSubClass         0 [unknown]
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x0bda Realtek Semiconductor Corp.
  idProduct          0xb832 802.11ac WLAN Adapter
  bcdDevice            0.00
  iManufacturer           1 Realtek
  iProduct                2 802.11ac WLAN Adapter

Btw. See how USB-idenfication is for 802.11ac. Confusing, that.

At this point, there is no pre-baked Linux kernel module. You can build one in a jiffy. Source code is at https://github.com/morrownr/rtl8852bu. Getting that C-code to work in your favor, after git clone, you need to run ./install-driver.sh. This script does a lot for you. At end, the script even stealthily copies the kernel module into proper directory to make loading the driver very easy. This is confusing and not all module builds do the install without asking.

When I modprobe 8852bu on Fedora 40, module does little bit of whining, but comes up. For the crash I sent some information to author: https://github.com/morrownr/rtl8852bu/issues/38

On my laptop, connection to 802.11ac / Wi-Fi 5 works ok. I suspect, there is something off with WPA3 as connections to 802.11ax / Wi-Fi 6 simply keep asking for network password in a forever loop. But hey! It worked. I got to The Net and was able to fix things. The thought of 802.11ax support is left unanswered. If USB-information doesn't state 802.11ax either, does the stick really support it or not? I dunno.

I'm hoping, WPA3-issue would be fixed one day, so that I'd be able to join any network of my choosing, not the one the device is capable of going.

Is that combo really impossible?

I've been running Linux in multiple old/oldish laptops. Broadcom is such a popular chip vendor, it is the most typical choice regardless of the laptop manufacturer. As Broadcom is full of idiots, their device drivers are proprietary. In their infinte wisdom, they pre-build Linux binaries and hand them out. This, obviously, is much better than not having the closed-source driver package at all. However, they really don't care. Every once in a while something happens in Linux kernel and their driver becomes less-than-operational.

Also, by lack of one official source, there are number of packages made out of the binary distribution. Key naming convention will include letters W and L in them, so you as an end user have to know wl stands for Broadcom BCM4311, BCM4312, BCM4313, BCM4321, BCM4322, BCM43224, BCM43225, BCM43227, BCM43228, BCM43142, BCM4331, BCM4352, BCM4360 devices. Simple! (No, not simple)
As an example: Fedora doesn't support Broadcom at all (as default), Arch Linux has packages brcm80211, b43 and broadcom-wl, Debian has  broadcom-wl as non-free package.

Recently my Fedora 40 refused to find a working Wifi. It all started on from 6.9.10 up to 6.10.5. Oh, I forgot to mention, altough Fedora doesn't have the support, there exist number of alternative RPM-repos for Fedora carrying Broadcom. An example, RPM Fusion., non-free package of broadcom-wl. To make this mess even messier, RPM Fusion also has kmod-wl and akmod-wl.

Ok, many packages, back to the problem: my laptop lost Wifi and I really struggled to figure out why, which kernel module caused the problem and for what reason. At the time both broadcom-wl and kmod-wl were installed.

This is VERY confusing! You have to be super-smart to understand much of my less-than-coherent story above. Unfortunately, that is the reality. Everything gets convoluted, confusing and chaos.

In RPM Fusion Bugzilla, there is a bug #6991 Kernel hangs due to broadcom wifi driver. This bug report is specifically for akmod-wl, which was not installed in my laptop. Using an USB-dongle, uninstalling both broadcom-wl and kmod-wl, followed by installing akmod-wl did not solve the problem either. Unlike with original packages, with akmod-wl there was no kernel crash on modprobe. With this package NetworkManager didn't work either. Weird.

When I wrote comments to bug report and Mr. Vieville, author of akmod-wl, replied with a suggestion. There existed an unreleased version of 6.30.223.271-53. Little bit of dnf installing and testing ... it works! Now my laptop had native Wifi-connectivity and I could un-plug the USB-dongle.

This incident left me really confused and happy.

I''ve been way too busy with my dayjob to do any blogging nor system maintenance.

Ever since S9y 2.4.0 update, my blog has been in a disarray. This has been a tough one for me as I'd love to run my system with much better quality.

Ultimately I had to find the required time and do tons of maintenance to get out of the sad state.

Mending activities completed:

• Changed hosting provider to Hetzner
• Rented a much beefier VM for the system
• Changed host CPU:
  • manufacturer from Intel to ?
  • architecture from AMD64 into ARMv8
  • generation from Pentium 4 into something very new
• Upgraded OS into Oracle Linux 9
• Upgraded DB into PostgreSQL 16
• Allocated more RAM for web server
• Tightened up security even more
• Made Google Search happier
• ... fixeds a ton of bugs in Brownpaper skin

Qualys SSL Report still ranks this site as A+ having HTTP/2. Netcraft Site report still ranks this site into better half of top-1-Million in The World.

Now everything should be so much better. Now this AI-generated image portrays a fixed computer system:

Twenty years with Fedora Linux. Release 39 is out!

I've ran Red Hat since 90s. Version 3.0.3 was released in 1996 and ever since I've had something from them running. As they scrapped Red Hat Linux in 2003 and went for Fedora Core / RHEL, I've had something from those two running. For the record: I hate those semi-working everything-done-only-half-way Debian/Ubuntu crap. When I want to toy around with something that doesn't quite fit, I choose Arch Linux.

To get your Fedora in-place-upgraded, process is rather simple. Docs are in article Performing system upgrade.

First you make sure existing system is upgraded to all latest stuff with a dnf --refresh upgrade. Then make sure upgrade tooling is installed: dnf install dnf-plugin-system-upgrade. Now you're good to go for download all the new packages with a: dnf system-upgrade download --releasever=39

Note: This workflow has existed for a long time. It is likely to work also in future. Next time all you have to do is replace the release vesion with the one you want to upgrade into.

Now all prep is done and you're good to go for the actual upgrade. To state the obvious: this is the dangerous part. Everything before this has been a warm-up run.

dnf system-upgrade reboot

If you have a console, you'll see the progress.

Installing the packages and rebooting into the new release shouldn't take too many minutes.

When back in, verify: cat /etc/fedora-release ; cat /proc/version
This resulted in my case:

Fedora release 39 (Thirty Nine)
Linux version 6.5.11-300.fc39.x86_64 (mockbuild@d23353abed4340e492bce6e111e27898) (gcc (GCC) 13.2.1 20231011 (Red Hat 13.2.1-4), GNU ld version 2.40-13.fc39) #1 SMP PREEMPT_DYNAMIC Wed Nov 8 22:37:57 UTC 2023

Finally, you can do optional clean-up: dnf system-upgrade clean && dnf clean packages

That's it! You're done for at least for next 6 months.

LM-Sensors is set of libraries and tools for accessing your Linux server's motheboard sensors. See more @ https://github.com/lm-sensors/lm-sensors.

If you're ever wondered why in Windows it is tricky to get readings from your CPU-fan rotation speed or core temperatures from you fancy GPU without manufacturer utilities. Obviously vendors do provide all the possible readings in their utilities, but people who would want to read, record and store the data for their own purposes, things get hairy. Nothing generic exists and for unknown reason, such API isn't even planned.

In Linux, The One toolkit to use is LM-Sensors. On kernel side, there exists The Linux Hardware Monitoring kernel API. For this stack to work, you also need a kernel module specific to your motherboard providing the requested sensor information via this beautiful API. It's also worth noting, your PC's hardware will have multiple sensors data providers. An incomplete list would include: motherboard, CPU, GPU, SSD, PSU, etc.

Now that sensors-detect found all your sensors, confirm sensors will output what you'd expect it to. In my case there was a major malfunction. On a boot, following thing happened when system started sensord (in case you didn't know, kernel-stuff can be read via dmesg):

systemd[1]: Starting lm_sensors.service - Hardware Monitoring Sensors...
kernel: nct6775: Enabling hardware monitor logical device mappings.
kernel: nct6775: Found NCT6793D or compatible chip at 0x2e:0x290
kernel: ACPI Warning: SystemIO range 0x0000000000000295-0x0000000000000296 conflicts with OpRegion 0x0000000000000290-0x0000000000000299 (_GPE.HWM) (20221020/utaddress-204)
kernel: ACPI: OSL: Resource conflict; ACPI support missing from driver?
systemd[1]: Finished lm_sensors.service - Hardware Monitoring Sensors.

This conflict resulted in no available mobo readings! NVMe, GPU and CPU-cores were ok, the part I was mostly looking for was fan RPMs and mobo temps just to verify my system health. No such joy. Uff.

It seems, this particular Linux kernel module has issues. Or another way to state it: mobo manufacturers have trouble implementing Nuvoton chip into their mobos. On Gentoo forums, there is a helpful thread: [solved] nct6775 monitoring driver conflicts with ACPI
Disclaimer: For ROG Maximus X Code -mobo adding acpi_enforce_resources=no into kernel parameters is the correct solution. Results will vary depending on what mobo you have.

Such ACPI-setting can be permanently enforced by first querying about the Linux kernel version being used (I run a Fedora): grubby --info=$(grubby --default-index). The resulting kernel version can be updated by: grubby --args="acpi_enforce_resources=no" --update-kernel DEFAULT. A reboot shows fix in effect, ACPI Warning is gone and mobo sensor data can be seen.

As a next step you'll need userland tooling to interpret the raw data into human-readable information with semantics. A new years back, I wrote about Improving Nuvoton NCT6776 lm_sensors output. It's mainly about bridging the flow of zeros and ones into something having meaning to humans. This is my LM-Sensors configuration for ROG Maximus X Code:

chip "nct6793-isa-0290"

    # 1. voltages
    ignore in0
    ignore in1
    ignore in2
    ignore in3
    ignore in4
    ignore in5
    ignore in6
    ignore in7
    ignore in8
    ignore in9
    ignore in10
    ignore in11

    label in12 "Cold Bug Killer"
    set in12_min 0.936
    set in12_max 2.613
    set in12_beep 1

    label in13 "DMI"
    set in13_min 0.550
    set in13_max 2.016
    set in13_beep 1

    ignore in14

    # 2. fans
    label fan1 "Chassis fan1"
    label fan2 "CPU fan"
    ignore fan3
    ignore fan4
    label fan5 "Ext fan?"

    # 3. temperatures
    label temp1 "MoBo"

    label temp2 "CPU"
    set temp2_max 90
    set temp2_beep 1

    ignore temp3
    ignore temp5
    ignore temp6
    ignore temp9
    ignore temp10
    ignore temp13
    ignore temp14
    ignore temp15
    ignore temp16
    ignore temp17
    ignore temp18

    # 4. other
    set beep_enable 1
    ignore intrusion0
    ignore intrusion1

I'd like to credit Mr. Peter Sulyok on his work about ASRock Z390 Taichi. This mobo happens to use the same Nuvoton NCT6793D -chip for LPC/eSPI SI/O (I have no idea what those acronyms are for, I just copy/pasted them from the chip data sheet). The configuration is in GitHub for everybody to see: https://github.com/petersulyok/asrock_z390_taichi

Also, I''d like to state my ignorance. After reading less than 500 pages of the NCT6793D data sheet, I have no idea what is:

• Cold Bug Killer voltage
• DMI voltage
• AUXTIN1 is or exactly what temperature measurement it serves
• PECI Agent 0 temperature
• PECI Agent 0 Calibration temperature

Remember, I did mention semantics. From sensors-command output I can read a reading, what it translates into, no idea! Luckily there are some of the readings seen are easy to understand and interpret. As an example, fan RPMs are really easy to measure by removing the fan from its connector. Here is an excerpt from my mobo manual to explain fan-connectors:

As data quality is taken care of and output is meaningful, next step is to start recording data. In LM-Sensors, there is sensord for that. It is a system service taking a snapshot (you can define the frequency) and storing it for later use. I'll enrich the stored data points with system load averages, this enables me to estimate a relation with high temperatures and/or fan RPMs with how much my system is working.

Finally, all data gathered into a RRDtool database can be easily visualized with rrdcgi into HTML + set of PNG-images to present a web page like this:

Nice!

For all those years I've been running SUSE-Linux, I've never bumped into this one while running a trivial zypper update:

Warning: File 'repomd.xml' from repository 'Update repository with updates from SUSE Linux Enterprise 15' is unsigned.

    Note: Signing data enables the recipient to verify that no modifications occurred
    after the data were signed. Accepting data with no, wrong or unknown signature can
    lead to a corrupted system and in extreme cases even to a system compromise.
    Note: File 'repomd.xml' is the repositories master index file. It ensures the
    integrity of the whole repo.

    Warning: We can't verify that no one meddled with this file, so it might not be
    trustworthy anymore! You should not continue unless you know it's safe.

continue? [yes/no] (no):

This error slash warning being weird and potentially dangerous, my obvious reaction was to hit ctrl-c and go investigate. First, my package verification mechanism should be intact and be able to verify if updates downloaded are unaltered or not. Second, there should have not been any breaking changes into my system, at least I didn't make any. As my system didn't seem to breached, I assumed a system malfunction and went investigating.

Quite soon, I learned this is less than rare event. It has happeded multiple times for other people. According to article Signature verification failed for file ‘repomd.xml’ from repository ‘openSUSE-Leap-42.2-Update’ there exists a simple fix.

By running two command zypper clean --all and zypper ref, the problem should dissolve.

Yes, that is the case. After a simple wash/clean/rinse -cycle zypper update worked again.
It was just weird to bump into that for the first time I'd assume this would have occurred some time earlier.

This is a deep dive into systems programing using Python. For those unfamiliar with programming, systems programming sits on top of hardware / electronics design, firmware programming and operating system programming. However, it is not applications programming which targets mostly end users. Systems programming targets the running system. Mr. Yadav of Dark Bears has an article Systems Programming is Hard to Do – But Somebody’s Got to Do it, where he describes the limitations and requirements of doing so.

Typically systems programming is done with C, C++, Perl or Bash. As Python is gaining popularity, I definitely want to take a swing at systems programming with Python. In fact, there aren''t many resources about the topic in the entire Internet.

Requirements

This is the list of basic requirements I have for a Python-based system daemon:

• Run as service: Must run as a Linux daemon, https://man7.org/linux/man-pages/man7/daemon.7.html
  • Start runing on system boot and stop running on system shutdown
• Modern: systemd-compatible, https://systemd.io/
  • Not interested in ancient SysV init support anymore, https://danielmiessler.com/study/the-difference-between-system-v-and-systemd/
• Modern: D-bus -connected
  • Service provided will have an interface on system D-bus, https://www.freedesktop.org/wiki/Software/dbus/
  • All Linux systems are built on top of D-bus, I absolutely want to be compatible
• Monitoring: Must support systemd watchdog, https://0pointer.de/blog/projects/watchdog.html
  • Surprisingly many out-of-box Linux daemons won't support this. This is most likely because they're still SysV-init based and haven't modernized their operation.
  • I most definitely want to have this feature!
• Security: Must use Linux capabilities to run only with necessary permissions, https://man7.org/linux/man-pages/man7/capabilities.7.html
• Security: Must support SElinux to run only with required permissions, https://github.com/SELinuxProject/selinux-notebook/blob/main/src/selinux_overview.md
• Isolation: Must be independent from system Python
  • venv
  • virtualenv
  • Any possible changes to system Python won't affect daemon or its dependencies at all.
• Modern: Asynchronous Python
  • Event-based is the key to success.
  • D-bus and systemd watchdog pretty much nail this. Absolutely must be asynchronous.
• Packaging: Installation from RPM-package
  • This is the only one I'll support for any forseeable future.
  • The package will contain all necessary parts, libraries and dependencies to run a self-contained daemon.

That's a tall order. Selecting only two or three of those are enough to add tons of complexity to my project. Also, I initially expected somebody else in The Net to be doing this same or something similar. Looks like I was wrong. Most systems programmers love sticking to their old habits and being at SysV-init state with their synchronous C / Perl -daemons.

Scope / Target daemon

I've previously blogged about running an own email server and fighting spam. Let's automate lot of those tasks and while automating, create a Maildir monitor of junk mail -folder.

This is the project I wrote for that purpose: Spammer Blocker

Toolkit will query for AS-number of spam-sending SMTP-server. Typically I'll copy/paste the IP-address from SpamCop's report and produce a CIDR-table for Postfix. The table will add headers to email to-be-stored so that possible Procmail / Maildrop can act on it, if so needed. As junk mail -folder is constantly monitored, any manually moved mail will be processed also.

Having these features bring your own Linux box spam handling capabilities pretty close to any of those free-but-spy-all -services commonly used by everybody.

Addressing the list of requirements

Let's take a peek on what I did to meet above requirements.

Systemd daemon

This is nearly trivial. See service definition in spammer-reporter.service.

What's in the file is your run-of-the-mill systemd service with appropriate unit, service and install -definitions. That triplet makes a Linux run a systemd service as a daemon.

Python venv isolation

For any Python-developer this is somewhat trivial. You create the environment box/jail, install requirements via setup.py and that's it. You're done. This same isolation mechanism will be used later for packaging and deploying the ready-made daemon into a system.

What's missing or to-do is start using a pyproject.toml. That is something I'm yet to learn. Obviously there is always something. Nobody, nowhere is "ready". Ever.

Asynchronous code

Talking to systemd watchdog and providing a service endpoint on system D-bus requires little bit effort. Read: lots of it.

To get a D-bus service properly running, I'll first become asynchronous. For that I'll initiate an event-loop dbus.mainloop.glib. As there are multiple options for event loop, that is the only one actually working. Majority of Python-code won't work with Glib, they need asyncio. For that I'll use asyncio_glib to pair the GLib's loop with asyncio. It took me while to learn and understand how to actually achieve that. When successfully done, everything needed will run in a single asynchronous event loop. Great success!

With solid foundation, As the main task I'll create an asynchronous task for monitoring filesystem changes and run it in a forever-loop. See inotify(7) for non-Python command. See asyncinotify-library for more details of the Pythonic version. What I'll be monitoring is users' Maildirs configured to receive junk/spam. When there is a change, a check is made to see if change is about new spam received.

For side tasks, there is a D-bus service provider. If daemon is running under systemd also the required watchdog -handler is attached to event loop as a periodic task. Out-of-box my service-definition will state max. 20 seconds between watchdog notifications (See service Type=notify and  WatchdogSec=20s). In daemon configuration file spammer-reporter.toml, I'll use 15 seconds as the interval. That 5 seconds should be plenty of headroom.

Documentation of systemd.service for WatchdogSec states following:

If the time between two such calls is larger than the configured time, then the service is placed in a failed state and it will be terminated

For any failed service, there is the obvious Restart=on-failure.

If for ANY possible reason, the process is stuck, systemd scaffolding will take control of the failure and act instantly. That's resiliency and self-healing in my books!

Security: Capabilities

My obvious choice would be to not run as root. As my main task is to provide a D-bus service while reading all users' mailboxes, there is absolutely no way of avoiding root permissions. Trust me! I tried everything I could find to grant nobody's process with enough permissions to do all that. No avail. See the documenatiion about UID / GID rationale.

As standard root will have waaaaaay too much power (especially if mis-applied) I'll take some of those away. There is a yet rarely used mechanism of capabilities(7) in a Linux. My documentation of what CapabilityBoundingSet=CAP_AUDIT_WRITE CAP_DAC_READ_SEARCH CAP_IPC_LOCK CAP_SYS_NICE means in system service definition is also in the source code. That set grants the process for permission to monitor, read and write any users' files.

There are couple of other super-user permissions left. Most not needed powers a regular root would have are stripped, tough. If there is a security leak and my daemon is used to do something funny, quite lot of the potential impact is already mitigated as the process isn't allowed to do much.

Security: SElinux

For even more improved security, I do define a policy in spammer-block_policy.te. All my systems are hardened and run SElinux in enforcing -mode. If something leaks, there is a limiting box in-place already. In 2014, I wrote a post about a security flaw with no impact on a SElinux-hardened system.

The policy will allow my daemon to:

• read and write FIFO-files
• create new unix-sockets
• use STDIN, STDOUT and STDERR streams
• read files in /etc/, note: read! not write
• read I18n (or internationalization) files on the system
• use capabilities
• use TCP and UDP -sockets
• acess D-bus -sockets in /run/
• access D-bus watchdog UDP-socket
• access user passwd information on the system via SSSd
• read and search user's home directories as mail is stored in them, note: not write
• send email via SMTPd
• create, write, read and delete temporary files in /tmp/

Above list is a comprehensive requirement of accesses in a system to meet the given task of monitoring received emails and act on determined junk/spam. As the policy is very carefully crafted not to allow any destruction, writing, deletion or manging outside /tmp/, in my thinking having such a hardening will make the daemon very secure.

Yes, in /tmp/, there is stuff that can be altered with potential security implications. First you have to access the process. While hacking the daemon, make sure to keep the event-loop running or systemd will zap the process within next 20 seconds or less. I really did consider quite a few scenarios if, and only if, something/somebody pops the cork on my daemon.

RPM Packaging

To wrap all of this into a nice packages, I'm using rpmenv. This toolkit will automatically wrap anything needed by the daemon into a nice virtualenv and deploy that to /usr/libexec/spammer-block/. See rpm.json for details.

SElinux-policy has an own spammer-block_policy_selinux.spec. Having these two in separate packages is mandatory as the mechanisms to build the boxes are completely different. Also, this is the typical approach on other pieces of software. Not everybody has strict requirements to harden their systems.

Where to place an entire virtualenv in a Linux? That one is a ball-buster. RPM Packaging Guide really doesn't say how to handle your Python-based system daemons. Remember? Up there ↑, I tried explaining how all of this is rather novel and there isn't much information in The Net regarding this. However, I found somebody asking What is the purpose of /usr/libexec? on Stackexchange and decided that libexec/ is fine for this purpose. I do install shell wrappers into /bin/ to make everybody's life easier. Having entire Python environment there wouldn't be sensible.

Final words

Only time will tell if I made the right design choices. I totally see Python and Rust -based deamons gaining popularity in the future. The obvious difference is Rust is a compiled language like Go, C and C++. Python isn't.

This topic won't go away. It just keeps bugging me. Back in -19 I wrote about GPE06 and couple months ago I wrote about sleep wake. As there is no real solution in existence and I've been using my Mac with Linux, I've come to a conclusion they are in fact the same problem.

When I boot my Mac, log into Linux and observe what's going on. Following CPU-hog can be observed in top:

   RES    SHR S  %CPU  %MEM   TIME+   COMMAND
     0      0 I  41.5   0.0   2:01.50 [kworker/0:1-kacpi_notify]

ACPI-notify will chomp quite a lot of CPU. As previously stated, all of this will go to zero if /sys/firmware/acpi/interrupts/gpe06 would be disabled. Note how GPE06 and ACPI are intertwined. They do have a cause and effect.

Also, doing what I suggested earlier to apply acpi=strict noapic kernel arguments:

grubby --args="acpi=strict noapic" --update-kernel=$(ls -t1 /boot/vmlinuz-*.x86_64 | head -1)

... will in fact reduce GPE06 interrupt storm quite a lot:

   RES    SHR S  %CPU  %MEM   TIME+   COMMAND
     0      0 I  10.0   0.0  0:22.92 [kworker/0:1-kacpi_notify]

Storm won't be removed, but drastically reduced. Also, the aluminium case of MBP will be a lot cooler.

However, by running grubby, the changes won't stick. Fedora User Docs, System Administrator’s Guide, Kernel, Module and Driver Configuration, Working with the GRUB 2 Boot Loader tells following:

To reflect the latest system boot options, the boot menu is rebuilt automatically when the kernel is updated or a new kernel is added.

Translation: When you'll install a new kernel. Whatever changes you did with grubby won't stick to the new one. To make things really stick, edit file /etc/default/grub and have line GRUB_CMDLINE_LINUX contain these ACPI-changes as before: acpi=strict noapic

Many people are suffering from this same issue. Example: Bug 98501 - [i915][HSW] ACPI GPE06 storm

Even this change won't fix the problem. Lot of CPU-resources are still wasted. When you close the lid for the first time and open it again, this GPE06-storm miraculously disappears. Also what will happen, your next lid open wake will take couple of minutes. It seems the entire Mac is stuck, but it necessarily isn't (sometimes it really is). It just takes a while for the hardware to wake up. Without noapic, it never will. Also, reading the Freedesktop bug-report, there is a hint of problem source being Intel i915 GPU.

Hopefully somebody would direct some development resources to this. Linux on a Mac hardware runs well, besides these sleep/sleep-wake issues.

Few years back I wrote about running Linux on a MacBook Pro. An year ago, the OpenSuse failed to boot on the Mac. Little bit of debugging, I realized the problem isn't in the hardware. That particular kernel update simply didn't work on that particular hardware anymore. Totally fair, who would be stupid enough to attempt using 8 years old laptop. Well, I do.

There aren't that many distros I use and I always wanted to see Fedora Workstation. It booted from USB and also, unlike OpenSuse Leap, it also booted installed. So, ever since I've been running a Fedora Workstation on encrypted root drive.

One glitch, though. It didn't always sleep wake. Quite easily, I found stories of a MBP not sleeping. Here's one: Macbook Pro doesn't suspend properly. Unlike that 2015 model, this 2013 puppy slept ok, but had such deep state, it had major trouble regaining consciousness. Pressing the power for 10 seconds obviously recycled power, but it always felt too much of a cannon for simple task.

Checking what ACPI has at /proc/acpi/wakeup:

Device  S-state   Status   Sysfs node
P0P2      S3    *enabled   pci:0000:00:01.0
PEG1      S3    *disabled
EC        S4    *disabled  platform:PNP0C09:00
GMUX      S3    *disabled  pnp:00:03
HDEF      S3    *disabled  pci:0000:00:1b.0
RP03      S3    *enabled   pci:0000:00:1c.2
ARPT      S4    *disabled  pci:0000:02:00.0
RP04      S3    *enabled   pci:0000:00:1c.3
RP05      S3    *enabled   pci:0000:00:1c.4
XHC1      S3    *enabled   pci:0000:00:14.0
ADP1      S4    *disabled  platform:ACPI0003:00
LID0      S4    *enabled   platform:PNP0C0D:00

For those had-to-sleep -cases, disabling XHC1 and LID0 did help, but made wakeup troublesome. While troubleshooting my issue, I did try if disabling XHC1 and/or LID0 would a difference. It didn't.

Also, I found it very difficult to find any detailed information on what those registered ACPI wakeup -sources translate into. Lid is kinda obvious, but rest remain relatively unknown.

While reading System Sleep States from Intel, a thought occurred to me. Let's make this one hibernate to see if that would work. Sleep semi-worked, but I wanted to see if hibernate was equally unreliable.

Going for systemctl hibernate didn't quite go as well as I expected. It simply resulted in an error of: Failed to hibernate system via logind: Not enough swap space for hibernation

With free, the point was made obvious:

               total        used        free      shared  buff/cache   available
Mem:         8038896     1632760     2424492     1149792     3981644     4994500
Swap:        8038396           0     8038396

For those not aware: Modern Linux systems don't have swap anymore. They have zram instead. If you're really interested, go study zram: Compressed RAM-based block devices.

To verify the previous, running zramctl displayed prettyy much the above information in form of:

NAME       ALGORITHM DISKSIZE DATA COMPR TOTAL STREAMS MOUNTPOINT
/dev/zram0 lzo-rle       7.7G   4K   80B   12K       8 [SWAP]

I finally gave up on that by bumping into article Supporting hibernation in Workstation ed., draft 3. It states following:

The Fedora Workstation working group recognizes hibernation can be useful, but due to impediments it's currently not practical to support it.

Ok ok ok. Got the point, no hibernate.

Looking into sleep wake issue more, I bumped into this thread Ubuntu Processor Power State Management. There a merited user Toz suggested following:

It may be a bit of a stretch, but can you give the following kernel parameters a try:

• acpi=strict
• noapic

I had attempted lots of options, that didn't sound that radical. Finding the active kernel file /boot/vmlinuz-5.18.18-200.fc36.x86_64, then adding mentioned kernel arguments to GRUB2 with: grubby --args=acpi=strict --args=noapic --update-kernel=vmlinuz-5.18.18-200.fc36.x86_64

... aaand a reboot!

To my surprise, it improved the situation. Closing the lid and opening it now works robust. However, that does not solve the problem where battery is nearly running out and I plug the Magsafe. Any power input to the system taints sleep and its back to deep freeze. I'm happy about the improvement, tough.

This is definitely a far fetch, but still: If you have an idea how to fix Linux sleep wake on an ancient Apple-hardware, drop me a comment. I'll be sure to test it out.

In ArchLinux, this is what happens too often when you're running simple upgrade with pacman -Syu:

error: libcap: signature from "-an-author-" is marginal trust
:: File /var/cache/pacman/pkg/-a-package-.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] y
error: failed to commit transaction (invalid or corrupted package)
Errors occurred, no packages were upgraded.

This error has occurred multiple times since ever and by googling, it has a simple solution. Looks like the solution went sour at some point. Deleting obscure directories and running pacman-key --init and pacman-key --populate archlinux won't do the trick. I tried that fix, multiple times. Exactly same error will be emitted.

Correct way of fixing the problem is running following sequence (as root):

paccache -ruk0
pacman -Syy archlinux-keyring
pacman-key --populate archlinux

Now you're good to go for pacman -Syu and enjoy upgraded packages.

Disclaimer:
I'll give you really good odds for above solution to go eventually rot. It does work at the time of writing with archlinux-keyring 20220713-2.

One of my boxes started failing after working pretty ok for years. On dnf update it simply said:

Curl error (6): Couldn't resolve host name for https://mirrors.fedoraproject.org/metalink?repo=fedora-35&arch=x86_64 [Could not resolve host: mirrors.fedoraproject.org]

Whoa! Where did that come from?

This is one of my test / devel / not-so-important boxes, so it might have been broken for a while. I simply haven't realized it ever happening.

Troubleshooting

Step 1: Making sure DNS works

Test: dig mirrors.fedoraproject.org

Result:

;; ANSWER SECTION:
mirrors.fedoraproject.org. 106 IN CNAME wildcard.fedoraproject.org.
wildcard.fedoraproject.org. 44 IN A 185.141.165.254
wildcard.fedoraproject.org. 44 IN A 152.19.134.142
wildcard.fedoraproject.org. 44 IN A 18.192.40.85
wildcard.fedoraproject.org. 44 IN A 152.19.134.198
wildcard.fedoraproject.org. 44 IN A 38.145.60.21
wildcard.fedoraproject.org. 44 IN A 18.133.140.134
wildcard.fedoraproject.org. 44 IN A 209.132.190.2
wildcard.fedoraproject.org. 44 IN A 18.159.254.57
wildcard.fedoraproject.org. 44 IN A 38.145.60.20
wildcard.fedoraproject.org. 44 IN A 85.236.55.6

Conclusion: Network works (obviously, I just SSHd into the box) and capability of doing DNS-requests works ok.

Step 2: What's in /etc/resolv.conf?

Test: cat /etc/resolv.conf

Result:

options edns0 trust-ad
; generated by /usr/sbin/dhclient-script
nameserver 185.12.64.1
nameserver 185.12.64.2

Test 2: ls -l /etc/resolv.conf

Result:

lrwxrwxrwx. 1 root root 39 Apr 25 20:31 /etc/resolv.conf -> ../run/systemd/resolve/stub-resolv.conf

Conclusion: Well, well. dhclient isn't supposed to overwrite /etc/resolv.conf if systemd-resolver is used. But is it?

Step 3: Is /etc/nsswitch.conf ok?

Previous one gave hint something was not ok. Just to cover all the bases, need to verify NS switch order.

Test: cat /etc/nsswitch.conf

Result:

# Generated by authselect on Wed Apr 29 05:44:18 2020
# Do not modify this file manually.

# If you want to make changes to nsswitch.conf please modify
# /etc/authselect/user-nsswitch.conf and run 'authselect apply-changes'.

d...

# In order of likelihood of use to accelerate lookup.
hosts:      files myhostname resolve [!UNAVAIL=return] dns

Conclusion: No issues there, all ok.

Step 4: Is systemd-resolved running?

Test: systemctl status systemd-resolved

Result:

● systemd-resolved.service - Network Name Resolution
     Loaded: loaded (/usr/lib/systemd/system/systemd-resolved.service; enabled; vendor pr>
     Active: active (running) since Mon 2022-04-25 20:15:59 CEST; 7min ago
       Docs: man:systemd-resolved.service(8)
             man:org.freedesktop.resolve1(5)
             https://www.freedesktop.org/wiki/Software/systemd/writing-network-configurat>
             https://www.freedesktop.org/wiki/Software/systemd/writing-resolver-clients
   Main PID: 6329 (systemd-resolve)
     Status: "Processing requests..."
      Tasks: 1 (limit: 2258)
     Memory: 8.5M
        CPU: 91ms
     CGroup: /system.slice/systemd-resolved.service
             └─6329 /usr/lib/systemd/systemd-resolved

Conclusion: No issues there, all ok.

Step 5: If DNS resolution is there, can systemd-resolved do any resolving?

Test: systemd-resolve --status

Result:

Global
       Protocols: LLMNR=resolve -mDNS -DNSOverTLS DNSSEC=no/unsupported
resolv.conf mode: stub

Link 2 (eth0)
Current Scopes: LLMNR/IPv4 LLMNR/IPv6
     Protocols: -DefaultRoute +LLMNR -mDNS -DNSOverTLS DNSSEC=no/unsupported

Test 2: resolvectl query mirrors.fedoraproject.org

Result:

mirrors.fedoraproject.org: resolve call failed: No appropriate name servers or networks for name found

Conclusion: Problem found! systemd-resolved and dhclient are disconnected.

Fix

Edit file /etc/systemd/resolved.conf, have it contain Hetzner recursive resolvers in a line:

DNS=185.12.64.1 185.12.64.2

Make changes effective: systemctl restart systemd-resolved

Test: resolvectl query mirrors.fedoraproject.org

As everything worked and correct result was returned, verify systemd-resolverd status: systemd-resolve --status

Result now:

Global
         Protocols: LLMNR=resolve -mDNS -DNSOverTLS DNSSEC=no/unsupported
  resolv.conf mode: stub
Current DNS Server: 185.12.64.1
       DNS Servers: 185.12.64.1 185.12.64.2

Link 2 (eth0)
Current Scopes: LLMNR/IPv4 LLMNR/IPv6
     Protocols: -DefaultRoute +LLMNR -mDNS -DNSOverTLS DNSSEC=no/unsupported

Success!

Why it broke? No idea.
I just hate these modern Linuxes where every single thing gets more layers of garbage added between the actual system and admin. Ufff.

Last November I created CentOS 8 -based Databricks containers.

At the time of tinkering with them, I failed to realize my base was off. I simply used the CentOS 8.4 image available at Docker Hub. On later inspection that was a failure. Even for 8.4, the image was old and was going to be EOLd soon after. Now that 31st Dec -21 had passed I couldn't get any security patches into my system. To put it midly: that's bad!

What I was supposed to be using, was the CentOS 8 stream image from quay.io. Initially my reaction was: "What's a quay.io? Why would I want to use that?"

Thanks Merriam-Webster for that, but it doesn't help.

On a closer look, it looks like all RedHat container -stuff is not at docker.io, they're in quay.io.

Simple thing: update the base image, rebuild all Databricks-images and done, right? Yup. Nope. The images built from steam didn't work anymore. Uff! They failed working that bad, not even Apache Spark driver was available. No querying driver logs for errors. A major fail, that!

Well. Seeing why driver won't work should be easy, just SSH into the driver an take a peek, right? The operation is documented by Microsoft at SSH to the cluster driver node. Well, no! According to me and couple of people asking questions like How to login SSH on Azure Databricks cluster, it is NOT possible to SSH into Azure Databricks node.

Looking at Azure Databricks architecture overview gave no clues on how to see inside of a node. I started to think nobody had ever done it. Also enabling diagnostic logging required the premium (high-prized) edition of Databricks, which wasn't available to me.

At this point I was in a full whatta-hell-is-going-on!? -mode.

Digging into documentation, I found out, it was possible to run a Cluster node initialization scripts, then I knew what to do next. As I knew it was possible to make requests into the Internet from a job running in a node, I could write an intialization script which during execution would dig me a SSH-tunnel from the node being initialized into something I would fully control. Obiviously I chose one of my own servers and from that SSH-tunneled back into the node's SSH-server. Double SSH, yes, but then I was able to get an interactive session into the well-protected node. An interactive session is what all bad people would want into any of the machines they'll crack into. Tunneling-credit to other people: quite a lot of my implementation details came from How does reverse SSH tunneling work?

To implement my plan, I crafted following cluster initialization script:

LOG_FILE="/dbfs/cluster-logs/$DB_CLUSTER_ID-init-$(date +"%F-%H:%M").log"
exec >> "$LOG_FILE"

echo "$(date +"%F %H:%M:%S") Setup SSH-tunnel"
mkdir -p /root/.ssh
cat > /root/.ssh/authorized_keys <<EOT
ecdsa-sha2-nistp521 AAAAE2V0bV+TrsFVcsA==
EOT

echo "$(date +"%F %H:%M:%S") Install and setup SSH"
dnf install openssh-server openssh-clients -y
/usr/libexec/openssh/sshd-keygen ecdsa
/usr/libexec/openssh/sshd-keygen rsa
/usr/libexec/openssh/sshd-keygen ed25519
/sbin/sshd

echo "$(date +"%F %H:%M:%S") - Add p-key"

cat > /root/.ssh/nobody_id_ecdsa <<EOT
-----BEGIN OPENSSH PRIVATE KEY-----
b3BlbnNzaC1rZXktdjEAAAAABG5vbmUAAAAEbm9uZQAAAAAAAAABAAAA
1zaGEyLW5pc3RwNTIxAAAACG5pc3RwNTIxAAAAhQQA2I7t7xx9R02QO2
rsLeYmp3X6X5qyprAGiMWM7SQrA1oFr8jae+Cqx7Fvi3xPKL/SoW1+l6
Zzc2hkQHZtNC5ocWNvZGVzaG9wLmZpAQIDBA==
-----END OPENSSH PRIVATE KEY-----
EOT
chmod go= /root/.ssh/nobody_id_ecdsa
echo "$(date +"%F %H:%M:%S") - SSH dir content:"

echo "$(date +"%F %H:%M:%S") Open SSH-tunnel"
ssh -f -N -T \
  -R22222:localhost:22 \
  -i /root/.ssh/nobody_id_ecdsa \
  -o StrictHostKeyChecking=no \
  nobody@my.own.box.example.com -p 443

Note: Above ECDSA-keys have been heavily shortened making them invalid. Don't copy passwords or keys from public Internet, generate your own secrets. Always! And if you're wondering, the original keys have been removed.

Note 2: My init-script writes log into DBFS, see exec >> "$LOG_FILE" about that.

My plan succeeded. I got in, did the snooping around and then it took couple minutes when Azure/Databrics -plumbing realized driver was dead, killed the node and retried the startup-sequence. Couple minutes was plenty of time to eyeball /databricks/spark/logs/ and /databricks/driver/logs/ and deduce what was going on and what was failing.

Looking at simplified Databricks (Apache Spark) architecture diagram:

Spark driver failed to start because it couldn't connect into cluster manager. Subsequently, cluster manager failed to start as ps-command wasn't available. It was in good old CentOS, but in base stream it was removed. As I got progress, also ip-command was needed. I added both and got the desired result: a working CentOS 8 stream Spark-cluster.

Notice how I'm specifying HTTPS-port (TCP/443) in the outgoing SSH-command (see: -p 443). In my attempts to get a session into the node, I deduced following:

As Databricks runs in it's own sandbox, also outgoing traffic is heavily firewalled. Any attempts to access SSH (TCP/22) are blocked. Both HTTP and HTTPS are known to work as exit ports, so I spoofed my SSHd there.

There are a number of different containers. To clarify which one to choose, I drew this diagram:

In my sparking, I'll need both Python and DBFS, so my choice is dbfsfuse. Most users would be happy with standard, but it only adds SSHd which is known not to work. ssh has the same exact problem. The reason for them to exist, is because in AWS SSHd does work. Among the changes from good old CentOS into stream is lacking FUSE. Old one had FUSE even in minimal, but not anymore. You can access DBFS only with dbfsfuse or standard from now on.

If you want to take my CentOS 8 brick-containers for a spin, they are still here: https://hub.docker.com/repository/docker/kingjatu/databricks, now they are maintained and get security patches too!

This is something Qualsys found: PwnKit: Local Privilege Escalation Vulnerability Discovered in polkit’s pkexec (CVE-2021-4034)

In your Linux there is sudo and su, but many won't realize you also have pkexec which does pretty much the same. More details about the vulnerability can be found at Vuldb.

I downloaded the proof-of-concept by BLASTY, but found it not working in any of my systems. I simply could not convert a nobody into a root.

Another one, https://github.com/arthepsy/CVE-2021-4034, states "verified on Debian 10 and CentOS 7."

Various errors I could get include a simple non-functionality:

[~] compile helper..
[~] maybe get shell now?
The value for environment variable XAUTHORITY contains suscipious content

This incident has been reported.

Or one which would stop to ask for root password:

[~] compile helper..
[~] maybe get shell now?
==== AUTHENTICATING FOR org.freedesktop.policykit.exec ====
Authentication is needed to run `GCONV_PATH=./lol' as the super user
Authenticating as: root

By hitting enter to the password-prompt following will happen:

Password:
polkit-agent-helper-1: pam_authenticate failed: Authentication failure
==== AUTHENTICATION FAILED ====
Error executing command as another user: Not authorized

This incident has been reported.

Or one which simply doesn't get the thing right:

[~] compile helper..
[~] maybe get shell now?
pkexec --version |
       --help |
       --disable-internal-agent |
       [--user username] [PROGRAM] [ARGUMENTS...]

See the pkexec manual page for more details.

Report bugs to: http://lists.freedesktop.org/mailman/listinfo/polkit-devel
polkit home page: <http://www.freedesktop.org/wiki/Software/polkit>

Maybe there is a PoC which would actually compromise one of the systems I have to test with. Or maybe not. Still, I'm disappointed.

In my previous post, I mentioned writing code into Databricks.

If you want to take a peek into my work, the Dockerfiles are at https://github.com/HQJaTu/containers/tree/centos8. My hope, obviously, is for Databricks to approve my PR and those CentOS images would be part of the actual source code bundle.

If you want to run my stuff, they're publicly available at https://hub.docker.com/r/kingjatu/databricks/tags.

To take my Python-container for a spin, pull it with:
docker pull kingjatu/databricks:python

And run with:
docker run --entrypoint bash

Databricks container really doesn't have an ENTRYPOINT in them. This atypical configuration is because Databricks runtime takes care of setting everything up and running the commands in the container.

As always, any feedback is appreciated.