Posts by BKFC

1) Questions and Answers : Unix/Linux : after Ubuntu update, BOINC is disconnected and unresponsive (Message 100767)
Posted 18 Mar 2021 by BKFC
Post:
SOLVED: The Ubuntu upgrade had reassigned a hard drive to another /dev slot.
2) Questions and Answers : Unix/Linux : after Ubuntu update, BOINC is disconnected and unresponsive (Message 100763)
Posted 18 Mar 2021 by BKFC
Post:
Correction: the machine is an AMD Ryzen 7 2700x, and accounts for about 90% of my Rosetta contribution (and occasionally a 5% badge).
3) Questions and Answers : Unix/Linux : after Ubuntu update, BOINC is disconnected and unresponsive (Message 100741)
Posted 15 Mar 2021 by BKFC
Post:
I just completed a minor update of Ubuntu on ARM machine. When I restarted, in the MATE interface, the BOINC icon was unresponsive. When I launched boincmgr from the command window, the BOINC window was blank, with the word 'disconnected' at the bottom. I then tried

% sudo /etc/init.d/boinc-client start
% sudo /etc/init.d/boinc-client status

from which I got this:

boinc-client.service - Berkeley Open Infrastructure Network Computing Client
Loaded: loaded (/lib/systemd/system/boinc-client.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Mon 2021-03-15 08:54:44 EDT; 14s ago
Docs: man:boinc(1)
Process: 18120 ExecStart=/usr/bin/boinc (code=exited, status=200/CHDIR)
Process: 18140 ExecStopPost=/bin/rm -f lockfile (code=exited, status=200/CHDIR)
Main PID: 18120 (code=exited, status=200/CHDIR)

Mar 15 08:54:44 Noether systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.
Mar 15 08:54:44 Noether systemd[18120]: boinc-client.service: Changing to the requested working directory failed: No such file or directory
Mar 15 08:54:44 Noether systemd[18120]: boinc-client.service: Failed at step CHDIR spawning /usr/bin/boinc: No such file or directory
Mar 15 08:54:44 Noether systemd[1]: boinc-client.service: Main process exited, code=exited, status=200/CHDIR
Mar 15 08:54:44 Noether systemd[18140]: boinc-client.service: Changing to the requested working directory failed: No such file or directory
Mar 15 08:54:44 Noether systemd[18140]: boinc-client.service: Failed at step CHDIR spawning /bin/rm: No such file or directory
Mar 15 08:54:44 Noether systemd[1]: boinc-client.service: Control process exited, code=exited, status=200/CHDIR
Mar 15 08:54:44 Noether systemd[1]: boinc-client.service: Failed with result 'exit-code'.

I really don't know how to fix this.
4) Message boards : Number crunching : BOINC: No protocol specified (Message 100185)
Posted 27 Dec 2020 by BKFC
Post:
Problem solved [apparently]: although one installs with the simple command "sudo apt install boinc", one must uninstall (by name) several components as in "sudo apt purge boinc boing-client boinc-manager". Once I did that, then reinstalled, then opened BOINC manager, I got the usual startup dialog box to choose a project.

So it appears to be running, except that R@H is not now sending tasks to any of my machines.

Thanks for your forbearance on this.
5) Message boards : Number crunching : BOINC: No protocol specified (Message 100146)
Posted 25 Dec 2020 by BKFC
Post:
I guess I can't convince anyone that I've installed this "from somewhere else." Be that as it may, I have shown what I have done to install/reinstall BOINC countless times on this machine. Reimaging Ubuntu from scratch just for BOINC is a non-starter: I have countless hours invested getting a variety of packages running for my own research.

I determine whether boinc is running or not via

/% ps -e | grep boinc

That said, looking at the systemctl status output, there is apparently a problem with the file gui_rpc_auth.cfg:

[Franklin]/etc/boinc-client/% !?stop?
sudo systemctl stop boinc-client.service
[Franklin]/etc/boinc-client/% !?start?
sudo systemctl start boinc-client.service
[Franklin]/etc/boinc-client/% !?stat?
sudo systemctl status boinc-client.service
● boinc-client.service - Berkeley Open Infrastructure Network Computing Client
Loaded: loaded (/lib/systemd/system/boinc-client.service; enabled; vendor preset: enabled)
Active: active (running) since Fri 2020-12-25 16:36:26 EST; 2s ago
Docs: man:boinc(1)
Main PID: 42915 (boinc)
Tasks: 2 (limit: 3833)
Memory: 3.1M
CGroup: /system.slice/boinc-client.service
└─42915 /usr/bin/boinc

Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] Setting up project and slot directories
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] Checking active tasks
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] Setting up GUI RPC socket
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] Can't open gui_rpc_auth.cfg - fix permissions
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] Checking presence of 0 project files
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 [---] This computer is not attached to any projects
Dec 25 16:36:26 Franklin boinc[42915]: 25-Dec-2020 16:36:26 Initialization completed
Dec 25 16:36:26 Franklin boinc[42915]: No protocol specified
Dec 25 16:36:27 Franklin boinc[42915]: Client is not authorized to connect to ServerNo protocol specified
Dec 25 16:36:28 Franklin boinc[42915]: Client is not authorized to connect to ServerNo protocol specified

I discovered that /var/lib/boinc-client/gui_rpc_auth.cfg is a soft link to etc/boinc-client/gui_rpc_auth.cfg, but the latter file does not exist. [I just did yet another purge/install to confirm this]

I have to stress here that I have not touched any of the /var/lib/boinc-clientor /etc/boinc-client files, except to delete all of them when I removed BOINC. If a file exists or not in those directory trees, it's a direct consequence of installation. If there's something wrong with permissions, I did not change them.
6) Message boards : Number crunching : BOINC: No protocol specified (Message 100125)
Posted 25 Dec 2020 by BKFC
Post:
I'm afraid I don't understand "running from your home directory". BOINC is not running in my home directory:

/% which boinc
/usr/bin/boinc

I've installed scores of packages on various Ubuntu machines (plus Raspberry Pi) via the following from my user prompt

/% sudo apt install [package name]

The result is an installation in various system-controlled directories, such as /etc and /var/lib. It does not install the application in my home directory.

There are a few files in my home directory that have to do with settings for the visual app. The critical files are in /etc/boinc-client or /var/lib/boinc-client.

I have tried this once more: I removed any file or directory in my home directory that had boinc or BOINC in its name, purged boinc via

/% sudo apt purge boinc

and then removed /etc/boinc-client and /var/lib/boinc-client

via rm -rf

I then installed boinc via

/% sudo apt install boinc

I then launched BOINC manager via a menu inside Ubuntu MATE. I could see that boinc and boincmgr were both running.

The menu interface in BOINC manager is completely greyed out, and the status "disconnected" at the bottom of the window.

As I have noted earlier, when I go through this process of installing BOINC for other machines (an 8-core AMD and Jetson Nano), I encounter none of these problems. The Odroid N2 is otherwise working: I can run C++ programs using OpenMP without any trouble.

If the solution to this is that I have to log into Ubuntu as root and install boinc, I have two problems:

1. It's different from every other apt install I've ever made.
2. I'm reluctant to have to make all preferences and adjustment always as root.

I believe BOINC has a conflict with Ubuntu 20.04 on ARM architecture. Repeatedly purging and reinstalling yields the same results.
7) Message boards : Number crunching : BOINC: No protocol specified (Message 100117)
Posted 23 Dec 2020 by BKFC
Post:
This all worked under 18.04...

I'm aware of the big/little distinction. When BOINC was working on 18.04, it was smart enough to assign a 2/3 numerical load to the big cores.

I always install via

/% sudo apt install boinc

I didn't know there were files in both /etc and /var/lib, so I have deleted all of them before reinstalling, which I have now done. It makes no difference. The machine is running headless, so I access it via a VPN client. When I launch BOINC manager, all options are greyed out and it says "disconnected" at the bottom of the window. When I try to connect it to the very same machine it's running on, it quickly reverts back to "disconnected." When I invoke 'top', both boinc and boincmgr are running.

Do I have to run this as user boinc? I did not with 18.04. There is also no boinc directory tree with the installation. I also run this on two other machines (a Ryzen7 8-core AMD, plus a Jetson Nano), installed in the same way, with no problems.
8) Message boards : Number crunching : BOINC: No protocol specified (Message 100104)
Posted 22 Dec 2020 by BKFC
Post:
I am now completely confused. Running 'top' shows 6 instances of R@H running, including one for over 100 hours, even though I have purged and reinstalled BOINC several times. I do not want R@H to run on all 6 cores, but all menu options in BOINC manager are grayed out, and the window still shows "disconnected" at the bottom. BOINC manager also lists no projects and no tasks, and /etc/boinc-client does not exist.

At this point I have no choice but to purge BOINC yet again, and delete the computer at the R@H site.
9) Message boards : Number crunching : BOINC: No protocol specified (Message 100099)
Posted 21 Dec 2020 by BKFC
Post:
I found a similar statement here:

Simply said, you have a permission problem when you get an error like this.
To fix this, you can either:
reinstall BOINC;
check that you're using the correct user to run BOINC.
slow down BOINC start at OS start-up, to allow all drivers to load first, see man sleep.

1. I've purged/reinstalled BOINC several times since installing Ubuntu 20.04.
2. I'm the only user.
3. I haven't started up the OS; there's no issue of slowing down BOINC.

Now I've also noticed that when BOINC manager actually starts inside Ubuntu-MATE, a line at the bottom of the window says "disconnected." When I try to connect using the BOINC password, it says "connecting..." then "disconnected" This is all on the same machine, not remote.

I'm coming to the conclusion that BOINC is incompatible with Ubuntu 20.04 running on an ARM platform.
10) Message boards : Number crunching : BOINC: No protocol specified (Message 100091)
Posted 21 Dec 2020 by BKFC
Post:
I had R@H running on an Odroid N2 with Ubuntu 18.04. After a clean install of 20.04, I have not been able to get it to run properly. Typically all processes are 'waiting' with nothing running. I then tried removing and reinstalling BOINC, but now I get the following:

/% boinc
21-Dec-2020 11:36:43 [---] cc_config.xml not found - using defaults
21-Dec-2020 11:36:43 [---] Starting BOINC client version 7.16.6 for aarch64-unknown-linux-gnu
21-Dec-2020 11:36:43 [---] Libraries: libcurl/7.68.0 OpenSSL/1.1.1f zlib/1.2.11 brotli/1.0.7 libidn2/2.2.0 libpsl/0.21.0 (+libidn2/2.2.0) libssh/0.9.3/openssl/zlib nghttp2/1.40.0 librtmp/2.3
21-Dec-2020 11:36:43 [---] Data directory: /home/bkeister
21-Dec-2020 11:36:43 [---] No usable GPUs found
21-Dec-2020 11:36:43 [---] Creating new client state file
21-Dec-2020 11:36:43 [---] libc: Ubuntu GLIBC 2.31-0ubuntu9.1 version 2.31
21-Dec-2020 11:36:43 [---] Host name: Franklin
21-Dec-2020 11:36:43 [---] Processor: 6 ARM Hardkernel ODROID-N2 [Impl 0x41 Arch 8 Variant 0x0 Part 0xd03 Rev 4]
21-Dec-2020 11:36:43 [---] Processor features: fp asimd evtstrm aes pmull sha1 sha2 crc32
21-Dec-2020 11:36:43 [---] OS: Linux Ubuntu: Ubuntu 20.04.1 LTS [4.9.236-106|libc 2.31 (Ubuntu GLIBC 2.31-0ubuntu9.1)]
21-Dec-2020 11:36:43 [---] Memory: 3.63 GB physical, 0 bytes virtual
21-Dec-2020 11:36:43 [---] Disk: 29.09 GB total, 18.01 GB free
21-Dec-2020 11:36:43 [---] Local time is UTC -5 hours
21-Dec-2020 11:36:43 [---] No general preferences found - using defaults
21-Dec-2020 11:36:43 [---] Preferences:
21-Dec-2020 11:36:43 [---] max memory usage when active: 1856.19 MB
21-Dec-2020 11:36:43 [---] max memory usage when idle: 3341.13 MB
21-Dec-2020 11:36:43 [---] max disk usage: 17.91 GB
21-Dec-2020 11:36:43 [---] don't use GPU while active
21-Dec-2020 11:36:43 [---] suspend work if non-BOINC CPU load exceeds 25%
21-Dec-2020 11:36:43 [---] (to change preferences, visit a project web site or select Preferences in the Manager)
21-Dec-2020 11:36:43 [---] Setting up project and slot directories
dir_open: Could not open directory 'slots' from '/home/bkeister'.
21-Dec-2020 11:36:43 [---] Checking active tasks
21-Dec-2020 11:36:43 [---] Setting up GUI RPC socket
21-Dec-2020 11:36:43 [---] Checking presence of 0 project files
21-Dec-2020 11:36:43 [---] This computer is not attached to any projects
21-Dec-2020 11:36:43 Initialization completed
No protocol specified
No protocol specified
No protocol specified
No protocol specified
No protocol specified
...etc

...where the "No protocol specified" goes on forever. Also, clicking on BOINC Manager in the GUI menu has no effect.

I don't know how to get past this.
11) Message boards : Number crunching : Another instance of BOINC is running (Message 99928)
Posted 5 Dec 2020 by BKFC
Post:
After several tries, the following seems to work for a mount point /aux on a HDD:

%/ sudo service boinc-client stop
%/ sudo cp -pR /var/lib/boinc-client /aux
%/ sudo rm -rf /var/lib/boinc-client
%/ sudo ln -sf /aux/boinc-client /var/lib/boinc-client
%/ sudo service boinc-client start
12) Message boards : Number crunching : How do I change the data directory in Ubuntu? (Message 99927)
Posted 5 Dec 2020 by BKFC
Post:
Those instructions are dated 2009, and among other things say to change the BOINC_DIR entry in /etc/defaults/boinc-client. I can confirm that boinc ignores that directive.

After several tries, the following seems to work for a mount point /aux on a HDD:

%/ sudo service boinc-client stop
%/ sudo cp -pR /var/lib/boinc-client /aux
%/ sudo rm -rf /var/lib/boinc-client
%/ sudo ln -sf /aux/boinc-client /var/lib/boinc-client
%/ sudo service boinc-client start
13) Message boards : Number crunching : Another instance of BOINC is running (Message 99917)
Posted 4 Dec 2020 by BKFC
Post:
It appears that the problem comes from the fact that I tried to move the data directory to another location (on an HDD instead of SSD). I have a separate query on how to do this but have no helpful responses so far. I have confirmed that changing the BOINC_DIR entry in /etc/default/boinc-client does nothing. Next I tried substituting a symlink for /var/lib/boinc-cilent. That's when all the trouble started. So I purged everything that had to do with 'boinc', including the symlink, and reinstalled. I no longer get the "another instance" failure message, but now all the files are back on my SSD. One would think this should be possible to move them elsewhere but I have found no recent documentation.
14) Message boards : Number crunching : Does boinc hold on to cores? (Message 99915)
Posted 4 Dec 2020 by BKFC
Post:
I had already selected that option, that is, unchecked the box.

I've run R@H for several months with no trouble. The only thing that has changed at my end is that I updated the NVIDIA driver to 4.50. There are some posts that suggest the problem is related to this, but I'm not in a position to downgrade my NVIDIA driver.
15) Message boards : Number crunching : How do I change the data directory in Ubuntu? (Message 99914)
Posted 4 Dec 2020 by BKFC
Post:
I tried that, and it did nothing. Then a later post (I can't find it now) said just to put in a symlink to where you want it. I also tried that and it didn't work, hence my query.
16) Message boards : Number crunching : Another instance of BOINC is running (Message 99913)
Posted 4 Dec 2020 by BKFC
Post:
I didn't tell it to start by other means. I found after a system restart that it wasn't running. When I did a restart it lasted a few seconds, then quit with the same error message.
17) Message boards : Number crunching : Another instance of BOINC is running (Message 99912)
Posted 4 Dec 2020 by BKFC
Post:
I just reinstalled everything yesterday before posting. I get it via apt get, and the version is 7.9.3.
18) Message boards : Number crunching : Another instance of BOINC is running (Message 99911)
Posted 4 Dec 2020 by BKFC
Post:
As noted in an earlier post, I have tried to manage this through sudo system boing-client restart [status], and I get the result that it exits with an error that another instance of BOINC is running, when in fact nothing is running at all.
19) Message boards : Number crunching : Another instance of BOINC is running (Message 99901)
Posted 4 Dec 2020 by BKFC
Post:
The tasks (one for the machine in question, one for a Jetson Nano which is very slow) may show up on the Rosetta web site, but at my end I have the following:

% ps -e | grep boinc
[nothing]

and

/% sudo service boinc-client status
● boinc-client.service - Berkeley Open Infrastructure Network Computing Client
Loaded: loaded (/lib/systemd/system/boinc-client.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Thu 2020-12-03 17:26:09 EST; 2h 41min ago
Docs: man:boinc(1)
Process: 3442 ExecStopPost=/bin/rm -f lockfile (code=exited, status=0/SUCCESS)
Process: 3426 ExecStart=/usr/bin/boinc (code=exited, status=108)
Main PID: 3426 (code=exited, status=108)

Dec 03 17:25:59 Noether systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.
Dec 03 17:26:09 Noether boinc[3426]: 03-Dec-2020 17:26:09 Another instance of BOINC is running.
Dec 03 17:26:09 Noether systemd[1]: boinc-client.service: Main process exited, code=exited, status=108/n/a
Dec 03 17:26:09 Noether systemd[1]: boinc-client.service: Failed with result 'exit-code'.

That is, boinc-client is not running.

If Rosetta is oversubscribed, that's a nice problem for Rosetta to have. Maybe I should look for another boinc application, though the 3 problems I have aren't Rosetta-specific.
20) Message boards : Number crunching : Another instance of BOINC is running (Message 99895)
Posted 3 Dec 2020 by BKFC
Post:
I now have another problem. I have uninstalled (via apt) and reinstalled boinc several times. When I launch, it runs for a few seconds and than quits:

%/ sudo service boinc-client status
● boinc-client.service - Berkeley Open Infrastructure Network Computing Client
Loaded: loaded (/lib/systemd/system/boinc-client.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Thu 2020-12-03 17:26:09 EST; 1s ago
Docs: man:boinc(1)
Process: 3442 ExecStopPost=/bin/rm -f lockfile (code=exited, status=0/SUCCESS)
Process: 3426 ExecStart=/usr/bin/boinc (code=exited, status=108)
Main PID: 3426 (code=exited, status=108)

Dec 03 17:25:59 Noether systemd[1]: Started Berkeley Open Infrastructure Network Computing Client.
Dec 03 17:26:09 Noether boinc[3426]: 03-Dec-2020 17:26:09 Another instance of BOINC is running.
Dec 03 17:26:09 Noether systemd[1]: boinc-client.service: Main process exited, code=exited, status=108/n/a
Dec 03 17:26:09 Noether systemd[1]: boinc-client.service: Failed with result 'exit-code'.

If I run 'ps' there is nothing there. There are various posts about lock files, but I don't know where to find them.

Until recently I've been able to run rosetta@home without a problem and have had either a 25% or 5% badge. But I have spent hours trying to get it to work, and at some point I have to devote my time (never mind the CPU time) to other tasks.


Next 20



©2024 University of Washington
https://www.bakerlab.org