After recent (L)Ubuntu (critical) system updates, the client refuses to start after required reboot. This is the 7.18.1 version, supplied by the Ubuntu update system. I have experienced this issue with 7.16.x too, but less frequent.

The only thing reported by the client is "status=192/n/a" and a random number of "dir_open: Could not open directory 'slot/xx' from '/var/lib/boinc-client'."
A huge number of empty "http_temp_xxxx" files are generated in /var/lib/boinc-client/. (File system is OK.) Log files, except for project related, are empty too.

So far, the only way I found around it, is to remove projects (WU's and all) one by one and test. Always ending up removing all of it. Reinstalling client and manager don't change anything. If I 'reset it' again, I will be loosing a number of completed and unreported WU's. (Yes, Collats Conjecture is down due to a physical move, due to several political issues ...). I have now two computers with this issue and one that I need to reboot (client still working) ...

Downloaded 7.4.22, but installing it creates a new work directory (not capable to handle finished WU's and in progress ...), so no go.

---

[ E5-2680v4 + i9-10900K + Q9550S + A72 + A57 + A55 + A53 + A15 + A7 + A5 + Kryo 465 | 629.5GB | RTX A2000 + Quadro M6000 + Tesla M40 + GTX 1650 + UHD Graphics 630 + Tegra X1 | Lubuntu 22.04/21.04/18.04 + Debian GNU/Linux 11 + Android 4-11 ]

Have you tried removing the lock file in your home directory?

Also can you confirm it is both client and manager not starting or just the manager? From time to time I get just the manager not starting and that always seems to be the lockfile problem.

Sounds to me like you lost permission to write to the data directory or files therein. BOINC Process exited with code 192 means it cannot write to the client_state.xml file, a process BOINC does upon starting up.
So it would seem that your OS updates remove (directory) permissions. Next time try to check who's owner of the files and folders before removing everything.

Looks like you are using the standard distro service installation. Try adding yourself to the boinc group.

First check to see who is in the boinc group

groups boinc

Then if you aren't in the boinc group, add yourself

sudo usermod -a -G boinc $(whoami)

Then see if you have access to the /var/lib/boinc-client directory

---

Yes, I started to remove those in the directories which the client was complaining about and later the rest. No change, except for complaining about other slots. (Interesting that lock files are used. I remember those from the early days of Windows, before M$ learned how to skip them.)

I always start the client using a terminal and got all the status from there. Next the manager. In this situation, the manager complaint about the client not running. Earlier times when this happened, I didn't have any unreported WU's and not that many in progress, so I uninstalled both the client and the manager, emptied the /var/lib/boinc-client/ directory and started over. Problem "solved" "quick 'n' dirty". This time, it's different ...

---

[ E5-2680v4 + i9-10900K + Q9550S + A72 + A57 + A55 + A53 + A15 + A7 + A5 + Kryo 465 | 629.5GB | RTX A2000 + Quadro M6000 + Tesla M40 + GTX 1650 + UHD Graphics 630 + Tegra X1 | Lubuntu 22.04/21.04/18.04 + Debian GNU/Linux 11 + Android 4-11 ]

OK, prior to my posting, I have been searching for the error code on the Internet and found nothing but other BOINC related codes. The directory, sub directories and files of concern have all the correct privileges (boinc/boinc), except for symlinks, but it shouldn't matter.

Interesting enough, the remaining computer which I have not rebooted yet have been somewhat problematic regarding filesystems. The system rejects write operations on a internal disk with NTFS, where I keep data used both in Linux and Windows. I even for awhile had a secondary disk with fs4, that was 'write protected'. Reinstalling the OS have not changed any thing. (I have experienced that Windows traditionally mess with hard drives that contain file systems other than those creates by M$. Windows ignores fs* and Debian/Ubuntu can use, but don't always behave properly with exFAT, according to own experience.) This computer behave nicely with Windows, so I can rule out hardware issues. All files and directories are set to root privileges for some unknown reason and it can not be changed, even logged in as root. Used as external drive with USB, no problems, but that's not a long term solution ... I have posted this issue on the Ubuntu forum several months ago, but no respond. One of the two computers where BOINC don't start had this issue once, but disappeared after reinstalling the OS.

Something I remembered from earlier occurrences. Every time BOINC failed to start after a system update (with required reboot), the Nvidia driver was updated to some degree. Now with driver v510 and earlier sometimes with v495, but not always. I doubt it relates, but I have noticed that Nvidia drivers has been picky when rearranging GPUs between my computers due to thermal issues. (L)Ubuntu may then end up in emergency mode due to failed video driver. (Never had this specific kind of issue with Windows. If eventually failed, Windows falls back to standard VGA resolution and it's an easy fix) Unfortunately, I'm not experienced enough with Linux yet to fix it, so it's much faster process to reinstall everything from scratch. This time, I would like to avoid that, due to all hours spent.

---

[ E5-2680v4 + i9-10900K + Q9550S + A72 + A57 + A55 + A53 + A15 + A7 + A5 + Kryo 465 | 629.5GB | RTX A2000 + Quadro M6000 + Tesla M40 + GTX 1650 + UHD Graphics 630 + Tegra X1 | Lubuntu 22.04/21.04/18.04 + Debian GNU/Linux 11 + Android 4-11 ]

To get Nvidia driver back without a reinstall just purge the Nvidia drivers and then reinstall them. When you purge them you automatically fall back to the Nouveau drivers after the rebooot. Then can install the Nvidia drivers again and reboot again to get them active.

sudo apt remove --purge *nvidia*
reboot
sudo apt install nvidia-driver-470 or
sudo apt install nvidia-driver-510
reboot

Both versions are in the current Debian and related distros.

If for some unknown reasons that the system doesn't automatically fall back to Nouveau, you can install them again with:

sudo apt install xserver-xorg-video-nouveau

---

Checked accordingly to your instructions and from the terminal it reported only 'boinc' and 'video render' as members. Added myself and still not a member. Checked via GUI and I was the only member.

---

[ E5-2680v4 + i9-10900K + Q9550S + A72 + A57 + A55 + A53 + A15 + A7 + A5 + Kryo 465 | 629.5GB | RTX A2000 + Quadro M6000 + Tesla M40 + GTX 1650 + UHD Graphics 630 + Tegra X1 | Lubuntu 22.04/21.04/18.04 + Debian GNU/Linux 11 + Android 4-11 ]

Thanks for the tip!

---

[ E5-2680v4 + i9-10900K + Q9550S + A72 + A57 + A55 + A53 + A15 + A7 + A5 + Kryo 465 | 629.5GB | RTX A2000 + Quadro M6000 + Tesla M40 + GTX 1650 + UHD Graphics 630 + Tegra X1 | Lubuntu 22.04/21.04/18.04 + Debian GNU/Linux 11 + Android 4-11 ]

There is no need to always update Ubuntu and variants.
I just recently updated my 18.04.1 to 18.04.4, which was mainly security updates, and these systems have been running this configuration for about 6 months now.
No need to fix what isn't broken, unless you are a daily user, and need the upgrades for what or what not.

In most cases, upgrades of the linux kernel will result in slower performance, as they get more and more bloated.
The only reason I can see why you would want the absolute latest, is if you're running the absolute latest CPUs (those that come with performance and efficiency cores).

There are AMD fixes for cpus in the latest releases that warrant upgrading the kernel for performance increases also.
Not just for Intel Alder Lake cpus.

---