You are not logged in.
Run alias
On my system the last line of output isalias ls='ls --color=auto'
When running as normal user, inside that vm, the output of alias was empty. Also no .bashrc under ~.
Did run dpkg -l inside and outside of vm, there were of course some differences. But not yet any clue, which missing or installed pkgs are the problem.
I guess i can create yet another vm and try to do something differently during install or during "install/remove packages after install"-phase. Like just install docker.io and docker-compose but remove nothing. (Avahi is considered a security risk and for me it's not absolutely necesssary to run cups inside or outside of vm and so on...)
On the other hand it would be useful to find out what i did wrong.
From the VirtialBox User Manual i found this:"Delayed writing through the host OS cache is less secure. When the guest OS writes data,
it considers the data written even though it has not yet arrived on a physical disk". That may explain the missing files and wrong content.
But i still think vbox could warn end-users about that, on the "Settings->Storage"-section.
I mean , sometimes vbox warns about wrong chipset or dislay settings. It could say something about that caching thing too...
Maybe i'll try that "hostiocache off"-thing. Hopefully usb keeps working without problems... or i have to start to use GitHub or Dropbox or something.
From topic https://dev1galaxy.org/viewtopic.php?pid=45777#p45777:
But another thing i noticed, with a new vm, docker-compose can't find the .yml-file with -f. I have to say it's absolute path, then d-c finds it.
Any ideas why? And how to fix that? I ask here, in the case this is a Devuan-specific problem. (Almost 20 years there were some kind of a RedHat-related problem with package gcc, that other distros did'nt have.)Something may have gone wrong with thet new vm's install. Or there is some important package missing. I noticed also that in the new vm's xfce terminal, there is no colours with output of "ls".
That's not weird when using root-account. But using a normal user account...
It is quite annoying, having to enter the full path for "-f"-option every time i run docker-compose inside that new vm. And the tab-completion with docker images doesn't work also.
Since there are no colors with "ls" inside xfce's terminal and most of the icons missing from desktop, i have a reason to believe that something went wrong with the install. Any ideas what could be missing?
It seems that i can make sqlalchemy fetch some data from tables when the connectionstrings are right.
But another thing i noticed, with a new vm, docker-compose can't find the .yml-file with -f. I have to say it's absolute path, then d-c finds it.
Any ideas why? And how to fix that? I ask here, in the case this is a Devuan-specific problem. (Almost 20 years there were some kind of a RedHat-related problem with package gcc, that other distros did'nt have.)
Something may have gone wrong with thet new vm's install. Or there is some important package missing. I noticed also that in the new vm's xfce terminal, there is no colours with output of "ls".
That's not weird when using root-account. But using a normal user account...
Hey, any ideas about editors that a) are available for Devuan and b) have an autosave function? I mean editors for coders....
I guess i can forget to save file when tired enough. Or save under a wrong name. But 2 or 3 weeks of commits missing from git's log? Come on.
When i shut down a vm by clicking it's window's close button i get options to a) save machine state b) send shutdown signal c) send power off. Anything but a) could undo the changes.
But I usually shut down vm by saying sudo /sbin/shutdown or sudo /sbin/halt inside tah vm. So vbox should commit changes into the virtual hd.
Has Oracle changed vbox's behaviour concerning shutting down a vm , after 6.0 ? Or is it just a bug that reveals itself in the situation where the backing fs is almost full ?
Managed to get rid of some complaints by changing "c.id_field==" to "c.columns.id_field==".
Maybe the software doesn't correspond to documentation or another way round.
Have to re-check that db-container works as it is supposed to. Tried with psql and got some errors.
Speaking of "doing non-default-things with vbox", i have tried virtual machine files on usb harddrive and virtual machine files on samba share, with just a .vbox-file on /home. It was a bit difficult but kind of worked. The most recent version of vbox then was 5.something.
I think running some light software inside vm wouldn't cause problems. I mean not eating too much storage or ram or processor capacity.
For some reason i would like to mention another non-default-thing i did with virtualbox 5.x. I tried doing dd if=/dev/usb-disk or dd of=/dev/usb-disk and managed to froze the host machine so badly that i had to turn it off by pushing powerbutton.
With vbox 6.0 that usb-situation got little better. Don't remember freezing my machine with dd after that.
I always manage to try Things That Are Forbidden.
... and the next funny thing was that usb devices stopped working properly. Vm couldn't find any usb stick at all although usb filters should "capture" all devices.
...and mouse not working either. Had to log out and in again to make usb work in the host side.
Today managed to reproduce the "mouse not working"-situation. Just add the usb mouse into vm's usb filters. And mouse pointer almost froze after that.
That may or may not explain why usb stick could not be found. A fair warning from vbox would be nice.
And that still wouldn't explain the missing files nor wrong content. Nor why there is 3 weeks of commits missing in git log. Didn't mention that detail week ago.
I still suspect that in some situations, vbox doesn't commit changes into virtual hd properly.
Maybe stackexchange.com would be a better place for this topic. But i don't have account on that site.
Anyways, changed logging settings in postgres-container. I started getting "LOG connection authetticated"-messages.
Entered into running python container with docker exec, started Python interpreter and started entering commands. Those "authenticated"-messages arrived into stderr. So now i know that sqlalchemy can connect into postgres.
But "r=conn.execute(text("SELECT * FROM table1"))", which i tried instead of "c.select().where()" is still a problem. For some reason it doesn't return any rows. I think it should.
And it is very stupid if i have to enter commands that way. I thought that with sqlalchemy, another way of doing queries would be possible.
Now i think that i have to confirm that postgres gets the sql commands at all. Parameter "echo=True" results in Python echoing the commands sqlalchemy generates. So i know that something happens on the Python-side of things.
But i'm not sure that those commands end up in postgres-container. So i'll have to tail -f some_log_file . Postres-container should echo some parts of the commands it receives, into stdout. And i haven't noticed any from that Flask-container.
Another thing is commands trying to select from database pg_catalog. I know these are necessary in order to copy pre-existing database structure into code.
The account i use to select from "table1" may not have necessary privileges to use that catalog. Have to do some gogling to find out which are the absolute minimal grants that i need.
IMO it is stupid to GRANT ALL PRIVILEGES ON ALL DATABASES just to select from 1 table.
But anyways, working example code would be very nice to find.
It was 2012 when i tried Flask before. I have the previous "hello_world"-attempt printed in paper. But it is half-finished and outdated. And then there is the principal problem that i don't need to define all the tables in code when i have already built the db, outside the Flask-container.
P.S. and then there is the question about psycopg2, which version actually works and can i download it from anywhere anymore? And is "no news is good news" still a valid principle?
By the way, any Python nerds in this forum? Experienced in sqlalchemy?
(This is Devuan-related in the sense that the host OS is Devuan Chimaera. Inside and outside the vm. ')
First time in 10 years i try to write a simple app with Flask and SQLAlchemy. Should be a simple thing, but more the instructions i read, the more confused i become. A three day parade march to Kabul is turning into a 10-year-long-nightmare-in-the-mountains.
I mean, it should't bee too difficult to:
1. set up connection to a pre-existing db
2. query a table with user-provided value
3. show result back to user somehow
But i have been running into many kinds of errors today.
I think setting up the conn works
e=sqlalchemy.create_engine(<A_CONNECTION_STRING>, echo=True)
conn=e.connect()
at least i get no complaints from Flask.
I query the table with:
metadata=MetaData()
c=Table('table1', metadata, autoload_with=e)
q=c.select().where(c.id_field==user_provided_id)
res=conn.execute(q).fetchone()
... and couldn't get to phase #3 . There were all kinds of complaints. For example:"AttributeError 'Table' object has no attribute 'id_field' ".
I have confirmed, with psql that i can connect to the db with the connectionstring(to be more precise:login, password and other params) i have. Also i have confirmed that name of the table and column are right. There also should be row with the id value i tried with.
Any working example code anaywhere which does what i'm looking for?
I mean i have read some of the sqlalchemy docs. Also tried https://www.datacamp.com/tutorial/sqlal … l-examples and https://www.blog.pythonlibrary.org/2010 … databases/ . To no avail.
One thing that can complicate simple things is Docker. I use bitnami/python:3.8.17. Tried also bitnami/python:latest. No difference, the Flask app is still not working.
Nothing surprising in Dockerfile: WORKDIR , COPY requirements.txt into container and RUN pip -r requirements.txt . And outside of Dockerfile, mount the app directory.
Inside vm, "dpkg -l docker*" says:
ii docker-compose 1.25.0-1
ii docker.io 20.10.5+dfsg10+deb11u2
And requirements.txt pip uses says, amongst other things:
Flask==3.0.0
psycopg2-binary==2.9.9
SQLAlchemy==2.0.22
There were also line: "flask-sqlalchemy==3.1.1". But i commented that out. To see if something changes. Nope. Not in a way that matters.
I suspect that those packages installed by pip also complicate a task that should be simple. New version of every package comes out 2 or 3 times a day and no one bothers to check if they work as intended or not. Which versions should work? Should i comment out flask-sqlalchemy or SQLAlchemy? Commenting out both will definitely prevent the db connection.
Somehow i'm not surprised that SQL injection is still a thing in 2020'ies. It seems that devs are forced to do queries in a very primitive and dangerous way.
Also it seems that it is mandatory to use scripting language abominations from the 1990'ies to get anything done at all.
Guess i'll look into that asound-thing some day. I may get bored booting with "super grub disk"-method, and kill 2 birds with 1 stone.
Not yet fully convinced that this sound problem is about configuration, but i'll find out soon enough.
https://www.youtube.com/watch?v=hJ-qU1nW0o0&t=0s Therion - Ginnungagap
not so much into Therion but this song sounds little like Bathory.
I had few installation attempts with sd cards. There were some problems. After that i have preferred optical discs as installation media.
Last weekend i tried chimaera netinstall on virtual machine. There were some problems also. I could install into a vm allright. But couldn't install any packages with apt. Not yet sure why. The usual suspects:
a) mis-configured apt
b) virtualbox problems i've been having lately
c) temporary network problems
d) something completely different.
I started using VBox with Squeeze if I remember correctly, and I haven't got any issues with "disk full" or access problems to USB drives.
Well, I have always avoided the default of using my home drive to store the VMs. From the beginning I have had a separate partition or drive for them. Its easy to configure in the options of the Virtualbox Manager.
Beside the point. Haven't yet confirmed what happens if /mnt/vbox runs out of storage but i wouldn't trust the .vdi-files to propertly store my files on that situation either.
Speaking of "doing non-default-things with vbox", i have tried virtual machine files on usb harddrive and virtual machine files on samba share, with just a .vbox-file on /home. It was a bit difficult but kind of worked. The most recent version of vbox then was 5.something.
I think running some light software inside vm wouldn't cause problems. I mean not eating too much storage or ram or processor capacity.
But couple of months a go it seemed like a good idea to install Docker inside a vm instead of outside. The iptables rules that Docker insists on wouldn't interfere with the iptables rules i wrote. And problems with dns and Docker would go away also. (Dockerd sees only the dns server that vbox runs, not dnsmasq or anything like that.)
On the other hand, you can fill up your /var very fast by pulling all kinds of images and building containers.
Maybe the usb-problem is not fully vbox's fault. Recently bought a new usb-mouse and found out that the older one co-operates little better with Fujitsu.
I also run apt-get install and apt-get upgrade -u recently, so some package may have broken usb somehow. On the other hand, problems with external mouse start after starting virtualbox.
also:
dmesg | grep -i snd
[ 65.384762] snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
[ 65.468638] snd_hda_codec_realtek hdaudioC0D0: ALC255: SKU not ready 0x909701f0
[ 65.468758] snd_hda_codec_realtek hdaudioC0D0: autoconfig for ALC255: line_outs=1 (0x14/0x0/0x0/0x0/0x0) type:speaker
[ 65.468760] snd_hda_codec_realtek hdaudioC0D0: speaker_outs=0 (0x0/0x0/0x0/0x0/0x0)
[ 65.468761] snd_hda_codec_realtek hdaudioC0D0: hp_outs=1 (0x21/0x0/0x0/0x0/0x0)
[ 65.468763] snd_hda_codec_realtek hdaudioC0D0: mono: mono_out=0x0
[ 65.468764] snd_hda_codec_realtek hdaudioC0D0: inputs:
maybe the second line is a good starting point, maybe not.
It has been a while since i had any audio-related problems with linux.
About 12 years ago, i upgraded Debian 4 to 5 or 6 and sometimes i couldn't get any sound out. I fixed that by saying "chmod o+rw /dev/dsp" as root.
Now it seems that there is no such thing as /dev/dsp anymore.
And the problem is, with this Fujitsu Lifebook U728 machine+Devuan Chimaera as OS,i can't get any sound out of it.
Audio mixer->Output Devices->show hardware outpur devices doesn't show any device at all.
Once played around with alsamixer. Didn't do anything at all.
Found this: https://bbs.archlinux.org/viewtopic.php?id=265211 , tried that "options snd-intel-dspcfg dsp_driver=1"-thing. Still hearing only the sound of silence.
Any ideas what to try next?
"lspci | grep -i audio" says:
00:1f.3 Audio device: Intel Corporation Sunrise Point-LP HD Audio (rev 21)
"lspci | grep -i sunrise" says also:
00:15.0 Signal processing controller: Intel Corporation Sunrise Point-LP Serial IO I2C Controller #0 (rev 21)
00:15.1 Signal processing controller: Intel Corporation Sunrise Point-LP Serial IO I2C Controller #1 (rev 21)
those lines may have something to do with that "dsp_driver=1"-thing.
"lsmod | grep snd" says:
snd_hda_codec_hdmi 73728 1
snd_hda_codec_realtek 167936 1
snd_hda_codec_generic 98304 1 snd_hda_codec_realtek
ledtrig_audio 16384 1 snd_hda_codec_generic
snd_soc_skl 180224 0
snd_soc_hdac_hda 24576 1 snd_soc_skl
snd_hda_ext_core 36864 2 snd_soc_hdac_hda,snd_soc_skl
snd_soc_sst_ipc 20480 1 snd_soc_skl
snd_soc_sst_dsp 36864 1 snd_soc_skl
snd_soc_acpi_intel_match 45056 1 snd_soc_skl
snd_soc_acpi 16384 2 snd_soc_acpi_intel_match,snd_soc_skl
snd_hda_intel 57344 0
snd_intel_dspcfg 28672 2 snd_hda_intel,snd_soc_skl
soundwire_intel 45056 1 snd_intel_dspcfg
snd_soc_core 319488 3 soundwire_intel,snd_soc_hdac_hda,snd_soc_skl
snd_compress 32768 1 snd_soc_core
snd_hda_codec 176128 5 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec_realtek,snd_soc_hdac_hda
snd_hda_core 110592 8 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_ext_core,snd_hda_codec,snd_hda_codec_realtek,snd_soc_hdac_hda,snd_soc_skl
snd_hwdep 20480 1 snd_hda_codec
snd_pcm 143360 8 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec,soundwire_intel,snd_compress,snd_soc_core,snd_soc_skl,snd_hda_core
snd_timer 49152 1 snd_pcm
snd 110592 10 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hwdep,snd_hda_intel,snd_hda_codec,snd_hda_codec_realtek,snd_timer,snd_compress,snd_soc_core,snd_pcm
soundcore 16384 1 snd
So, any ideas how to make xfce's audio mixer find that sunrise audio device?
Quick solution would be to buy an usb soud card. I know one card that Just Works or at least it used to.
But it would be better to make the internal sound chip work.
I think it is too buggy for "production use".At least that version which i downloaded from virtualbox.org's repository.
In last 4 months i have encountered 2 times "VDI_DISK_FULL"-complaint.
The second time it became clear to me that cause was /home running out of space.
One nice thing related to that:at least when /home is full or almost full, i cannot trust virtual harddrives of vbox.
First found out that there was some files missing inside the vm. Copied them from another vm's vhd.
Next thing i found out that the contents were wrong. I know i commented out that part of script but now i have to comment that out again. And try to remember what other changes i have made after commenting out that particular part.
Luckily i have made backups into usb stick.
... and the next funny thing was that usb devices stopped working properly. Vm couldn't find any usb stick at all although usb filters should "capture" all devices.
...and mouse not working either. Had to log out and in again to make usb work in the host side.
Maybe it is just this particular hw+sw that is the problem. Unfortunately i have only 1 computer with recent enough processor to run vbox.
I'll try 6.1 if that co-operates better. If that doesn't work, i'll have to move to Xen or something.
I was hoping that i don't have to make backups or restore from hdd again. But i can do it if i really have to.
Btw, one thing i noticed:audio is not working. Tried some youtube videos and i can't hear a thing. Maybe it is just about volume, maybe about something else. Any ideas how to fix that?
Edit. found this:https://bbs.archlinux.org/viewtopic.php?id=265211 . I'll try that "options snd-intel-dspcfg dsp_driver=1"-thing soon.
Hi rolfie, a quick question please.
Is this specific to the .deb file downloaded from the virtualbox site? I use the virtualbox packages from debian - fast track repo for chimaera, daedalus, and excalibur - and the regular repo for ceres.
It's tricky when you install debian's packages into devuan. Might work, but might not also. Nice to know that in this case, vbox could work.
Complaints about usb enumeration went away after upgrading the extpack. Tried the vm once, it was a bit slow to notice, but finally cat /proc/partitions showed /dev/disk/by-uuid/usb-stick.
The essential parts of the problem are in the subject.
Almost 4 months ago i installed virtualbox 7 by adding "deb [arch=amd64 signed-by=/usr/share/keyrings/oracle-virtualbox-2016.gpg] https://download.virtualbox.org/virtualbox/debian bullseye contrib" . Also installed then-current extension pack from virtualbox.org . /sbin/vboxconfig and update-initramfs also done. Vbox found memory stick allright. Back in the day when my "new" desktop was still working.
(Maybe installing version 6.x would have been a better idea, i think.)
Now i had to start using Fujitsu. Copied partitions into ssd from hdd by using dd.
I have done that "/sbin/vboxconfig" -trick again. Also update-initramfs and update-grub. Still, with Fujitsu, vbox can't enumerate usb devices. And inside vm i cannot find the attached memorystick. Lsusb cannot find it, nor can cat /proc/partitions.
Any ideas what to try next?
I have recently run apt-get upgrade -u and downloaded also the most recent version of extension pack for installing. Will install the new one soon.
How about users and groups? Should the normal_user belong to some group in order to be able to use usb-devices?
Again, i wasn't precise enough with my previous message.
Not planning to make a mixture of new and old /.
The thing is:I don't want do destroy or format my /home-partition on Fujitsu. It will take some time to copy it back from hdd via usb.
I also need to keep the old / around long enough to copy some necessary files from /etc. Now i think that i got the contents for most of those files right, abt 4 months ago.
On the other hand, i think that configuration files i extracted from a memory stick are mostly unusable rotten crap.
Also root on lvm for new / would be nice but not absolutely necessary. Although no need to destroy previous vg containing /home for that.
So, bypassing the partitioning-part in installation process seems like a good idea.
Btw. should it be possible to install Devuan on a pre-existing partition?
I prefer not to erase/delete/format the / or /home i have copied into Fujitsu's ssd. I know that with Debian there at least was something called "debootstrap". I remember that with that there was an option to skip the partitioning/formatting-parts of the installation process.
Can refractainstaller do that?