hdparm APM level when AC adapter is plugged/unplugged
Hi, It seems /usr/lib/pm-utils/power.d/95hdparm-apm script from the hdparm package may be called when AC power adapter is plugged or unplugged (true/false arguments). I see that this script is invoked by /lib/udev/rules.d/85-hdparm.rules through /lib/udev/hdparm on boot or resume. What package should be installed to call the script in response of change or power supply? Is that part of the script survived from the times of HAL? I expect some udev rule that acts for SUBSYSTEM=power_supply. Perhaps laptop-mode-tools may be installed as an alternative solution and I will have once more tool that does something with disks. My worries are that it may interact with hdparm and udisks2 in some strange way. The context of the question is the following. I added a hard disk drive and before I have created partitions I noticed regular sound of parking heads. It is udisks2 that decided to check SMART every 10 minutes for the disk with no mounted partition. The Toshiba disk reports 254 APM level (likely set by hdparm), active/idle state and parks heads with 2minutes timeout. I have not managed to read value of this timeout using hdparm or sdparm. As to udisks2, it is not possible to change polling interval, but likely it can be instructed to ignore particular disk without disabling SMART completely. The price is that udisks will ignore the disk completely. If I explicitly set standby (spindown) timeout <10min (e.g. by creating /etc/udisks2/.conf file) then disk is not waken up from standby state by udisks2 any more. However hdparm may distinguish on battery and connected to mains states, so it is more flexible than udisks. Since I have dug into this problem, I decided to check whether disk APM value is adjusted when power state is changed after resume (during resume 128/254 values are chosen depending on on_ac_power script). So I am disappointed by default behavior of tools, but I hope that I forgot to install some package that calls scripts from hdparm on connect/disconnect of power adapter.
Re: Color of the active window title bar in ubuntu-mate?
On 22/08/2022 18:02, Greg Wooledge wrote: On Mon, Aug 22, 2022 at 10:02:22AM +, Victor Sudakov wrote: Any applications using standard decorations (Firefox, MATE Terminal, Google Chrome etc). It's worth pointing out, perhaps, that Google Chrome does *not* use the standard window manager decorations by default. There is, however, an option you can toggle to make it do so. (Unless this behavior changed in some recent version -- but I doubt it.) Firefox, on the other hand, *does* use the regular widgets. Behavior of Firefox depends on GTK_CSD and XDG_CURRENT_DESKTOP environment variables. Some time ago there were more if's for Gnome, KDE, etc.: - https://hg.mozilla.org/integration/autoland/file/d58f8264/widget/gtk/nsWindow.cpp#l9192 - https://hg.mozilla.org/integration/autoland/file/tip/widget/gtk/nsWindow.cpp#l9192 The latter link is for current version without fixed revision. "Customize toolbar" dialogue has "Titlebar" checkbox to change behavior. Recently I faced an issue with fluxbox and I blamed client side decoration at first, so I tried to disable it.
Re: failing HDD, ddrescue says remaning time is 7104d
On 01/09/2022 06:33, David Christensen wrote: First sentence -- You don't want the OS to access the drive on the next boot. I would say that while ddrescue is running any other disk operations, especially causing seeks, may be undesired. While smartd may be configured or stopped, udisksd is tightly integrated into desktop environments and 10 min SMART polling interval is hardcoded for SATA disks (however USB enclosures are ignored). https://github.com/storaged-project/udisks/issues/407 In some cases it still might be better to monitor disk temperature. Unfortunately I can not suggest anything concerning too low read speed, I do not have such experience. It might be problem with particular area on disk surface, so speed should increase for other block ranges, or it might be firmware failure. In the case of really precious data try to find a service with appropriate equipment. Otherwise attempt of reading may cause so severe damage that nobody will be able to recover any information.
Re: question re tar
On 19/09/2022 02:37, jr wrote: when I saw the links and started investigating, I tried cat for the names, ie $ tar -cvWf $arcname $(cat $fnames) adding one or two file names on the command line works as expected, supplying names from list and or file produces those links. It seems some files are present multiple times in your list. echo text >file.txt tar cvWf test.tar file.txt file.txt tar tvf test.tar -rw-rw-r-- user/user 5 2022-09-20 17:43 file.txt hrw-rw-r-- user/user 0 2022-09-20 17:43 file.txt link to file.txt
Re: question re tar
On 22/09/2022 14:00, jr wrote: On Wednesday, 21 September 2022 at 17:20:05 UTC+1, Markus Schönhaber wrote: Could you please stop using a mail client that starts a new thread with every message you send? Please use something instead that really creates a reply when you are replying to someone (i. e. something that sets the In-Reply-To/References message headers accordingly). ouch. I read the digest, and in that it appears as a single thread. and I've just checked 'linux.debian.user' on Google Groups, there too I see a single thread only. cannot see "a new thread [started] with every message", sorry. Debian mail list archive has a rare mhonarc configuration that adds reply to list action (usually only reply to sender is available) and these mailto: links contain proper In-Reply-To value https://lists.debian.org/debian-user/2022/09/msg00584.html Even groups.google.com is unable to render proper discussion tree, its representation is a list. Mail user agents may intentionally suppress threads based on heuristics to avoid combining unrelated discussions having subjects like "question" or "problem".
Re: mail headers not set (cf question re tar)
On 24/09/2022 02:37, jr wrote: On Thursday, 22 September 2022 at 16:30:05 UTC+1, Max Nikulin wrote: ... Debian mail list archive has a rare mhonarc configuration that adds reply to list action (usually only reply to sender is available) and these mailto: links contain proper In-Reply-To value https://lists.debian.org/debian-user/2022/09/msg00584.html Even groups.google.com is unable to render proper discussion tree, its representation is a list. Mail user agents may intentionally suppress threads based on heuristics to avoid combining unrelated discussions having subjects like "question" or "problem". thanking all for the observations + information. I have had a look in settings of Gmail, my "mua", and cannot find an obvious (any!) place where headers could be set (and haven't the knowledge, anyhow). so, I won't try and start any more threads for now (for which sorry) and hope that someone can/will supply Gmail specific instructions "how to"; to that end please write an email with 'mail headers' in the subject, thank you. Sorry, I rarely use gmail web UI, in my case "mailto:"; links are handled by thunderbird. If the browser app is the only mail user agent available on your chromebook and it can not handle In-Reply-To link parameter then it is sour. Since you mentioned groups.google.com, does it set proper headers when reply to particular message is sent using this service? However it looks like a kind vendor lock to their own mail list archive due to unsatisfactory support of primary external archives.
Re: Some applications don't follow default browser settings
On 04/10/2022 11:07, Tianyu Chen wrote: I'm using Google Chrome (/opt/google/chrome/google-chrome) as my default browser, which is set by `update-alternatives` command. However my firefox-esr is always launching when links are opened in Telegram and Mozilla Thunderbird. Almost certainly Thunderbird ignores alternatives, mailcap and queries XDG settings from mimeapps.list, mimeinfo.cache files $ xdg-settings get default-web-browser google-chrome.desktop xdg-mime query default x-scheme-handler/http xdg-mime query default x-scheme-handler/https or xdg-settings get default-url-scheme-handler http that likely invokes xdg-mime $ cat ~/.config/mimeapps.list [Default Applications] x-scheme-handler/tg=userapp-Telegram Desktop-ZS5OS1.desktop text/html=google-chrome.desktop My guess is that when applications are going to open a link, MIME type is not known yet, so scheme handler is queried. Likely you miss the following associations in the [Default Applications] section: x-scheme-handler/http=google-chrome.desktop; x-scheme-handler/https=google-chrome.desktop; Do not forget to run update-desktop-database with a proper directory as its argument.
Re: Can't open email attachments with Firefox 102.3.0 esr
On 05/10/2022 16:09, Mick Ab wrote: Firefox was updated to Firefox 102.3.0 esr (64 bit). It has subsequently been found that this version of Firefox does not allow email attachments to be opened. Clicking on such an attachment used to always open a dialogue box which had options "Open with..." and "Save" but now it just jumps straight to a Download box with a list of folders to download to. Handling of downloads changed in Firefox since version 91 https://support.mozilla.org/en-US/kb/manage-downloads-preferences-using-downloads-menu How file downloads are handled in Firefox version 98 and above https://support.mozilla.org/en-US/kb/where-find-and-manage-downloaded-files-firefox Where to find and manage downloaded files in Firefox This change caused hot discussions in their bug tracker and issues like https://bugzilla.mozilla.org/1738574 Want an option to continue to use /tmp and autodelete downloads automatically opened in an application As a temporary measure I set browser.download.improvements_to_download_panel to false, but was struggling with another problem. Some other recipe may be more suitable for you as a long term solution.
Re: Cannot read newsgroups with new Thunderbird
On 05/10/2022 20:08, Miroslav Skoric wrote: After a recent Thunderbird upgrade in Buster (from version 91-something to 101-something, or like), it stopped handling newsgroups properly (where the source is News Server (NNTP) on the same machine, and there nothing was changed/upgraded). If you can provide steps to reproduce then it may be reasonable to report the issues to https://bugzilla.mozilla.org Likely you will be asked to confirm the problem for the latest release or daily build. A part of NNTP has been reimplemented since Thunderbird-91, I am unsure concerning Kamil's suggestion to switch back to the older one. There are enough bugs in the latter as well. On the other hand my issues was related to network latency or lost packets. In some aspects behavior of Thunderbird-102 should be better, but I am not surprised by new bugs.
Re: sigc++ library missing object.h file
On 08/10/2022 04:54, Gary L. Roach wrote: I've been trying to compile an older program from source code and keep getting the error missing sigc++/object.h file. .. I really need this program (gspeaker). Removing the # from the offending module only caused a slew of other errors. What would be in this object.h file? Have you tried any general purpose search engine? I was curious if it is related somehow to SGI. The result of my attempt: https://github.com/libsigcplusplus/libsigcplusplus/blob/master/NEWS 2.5.2 (unstable): * Remove useless headers: sigc++/class_slot.h sigc++/hide.h sigc++/method_slot.h sigc++/object.h sigc++/object_slot.h sigc++/retype.h Some of these still have equivalents in sigc++/adaptors/ (Kjell Ahlstedt, Murray Cumming) Bug #752560 Perhaps it is possible to modernize source code of your precious program. Read that "other" errors, library docs and maybe git history.
Re: Cannot read newsgroups with new Thunderbird
On 07/10/2022 01:14, piorunz wrote: On 06/10/2022 00:05, Ralph Katz wrote: news://news.gmane.io/gmane.linux.debian.user and replying directly to debian-user. Wow! Is this free service? The story of gmane: https://lars.ingebrigtsen.no/2020/01/15/news-gmane-org-is-now-news-gmane-io/
Re: sigc++ library missing object.h file
On 08/10/2022 23:17, Gary L. Roach wrote: Thank you all for your replies, I mean gspeaker not gespeaker. The first is a speaker enclosure design program and the second is a front end for the sound system. Not the same -- damn, I had hopes. I also went back and looked at the older versions of sigc++ and found nothing. I posted a link to version control system expecting that it is possible to check content of the header in old versions. Perhaps by comparing declarations and includes there with current headers and docs some person from the program community may fix compilation errors by specifying proper includes.
Re: Cheap NAS
On 18/10/2022 08:59, Stefan Monnier wrote: - In my experience, whether `hdparm -S..` works on a USB drive depends on the USB<->SATA adapter in use. If it doesn't, `hdparm` will tell you, tho. I did go further but my impression was that in the case of USB disk enclosures sdparm may work with power saving parameters inaccessible to hdparm. - On my WD drives, the `-S` seems to be completely ignored and instead the spin down time is chosen by the drive based on the APM setting (i.e. `hdparm -B...`). I heard there's a `idle3-tools` thingy that may be able to provide finer control, but I never tried it. I believed that idle3-tools is for Seagate. Have it got support for WD as well? I do not mind to have a proper tool for Toshiba. What prevents me from starting experiments is that I am afraid that a disk may be transformed into a brick.
Re: Cheap NAS
On 17/10/2022 22:50, Tixy wrote: # Set spindown time for disk if NAS1=`findmnt -n -o SOURCE /nas1/main`; then /sbin/hdparm -S120 $NAS1; fi Rather than hard code the disk as /dev/sdc I use 'findmnt' to get the disk name from the point it's mounted at (/nas1/main in this case). Are there known problems with using e.g. WWN (/dev/disk/by-id/... in /etc/hdparm.conf) for per-disk configuration? I expect that should avoid issues with changing /dev/sdX devices.
Re: grep replacement using sed is behaving oddly
On 22/10/2022 02:09, The Wanderer wrote: 'info sed', section 'sed regular expressions', subsection 'regular expression extensions': While a reader may find more interesting stuff lying around while traveling by this path, there is a shorthand info "(sed) regexp extensions" and alternatives that might be more friendly: - infobrowser "(sed) regexp extensions" (configurable through update-alternatives) - tkinfo "(sed) regexp extensions" - emacs -f info-standalone "(sed) regexp extensions" or M-: (info "(sed) regexp extensions") RET within a running emacs instance https://www.gnu.org/software/sed/manual/html_node/regexp-extensions.html https://www.gnu.org/software/sed/manual/sed.html#regexp-extensions
Re: grep replacement using sed is behaving oddly
On 22/10/2022 20:23, Gary Dale wrote: sed -i '//d' *.html did the trick. I would suggest you to use more specific pattern to avoid removing of meaningful text due to a lost newline character: sed -i -e '/^\s*]*>\s*$/d' "." in regexp may be a source of surprises (or catastrophic backtracking in more complex regexps) printf 'A bc\n' | sed -e 's/<.*>/\n/' A c printf 'A bc\n' | sed -e 's/<.[^>]>/\n/g' A b c
Re: Remove Network Manager panel icon in Mate 1.24.1 Desktop
On 25/10/2022 15:50, jeremy ardley wrote: My only problem now is the annoying NetworkManager icon in the Mate panel. I want to remove it and I have gone through various advice pages but nothing seems to work. The NetworkManager icon seems permanently wedged into the panel. Something like /etc/xdg/autostart/nm-applet.desktop might be the reason. I am unsure though. Mate may use another GUI.
Re: How to can make a partition in my hard disk ?
On 26/10/2022 04:34, William Torrez Corea wrote: Model Family: Toshiba L200 (SMR) I would not consider a HDD with shingled magnetic recording as a storage to install OS. This series has a CMR product, but it has another problem: 3 Spin_Up_TimePOS--K 100 100 001-1318 4 Start_Stop_Count-O--CK 100 100 000-606 193 Load_Cycle_Count-O--CK 093 093 000-75472 It has quite aggressive policy to unload heads. It happens approximately than once per minute. Following read operations may have noticeable delay. At first I suspected delays due time required to spin up stopped disk, but in your case Power_Cycle_Count = Start_Stop_Count, so your system keeps the disk spinning. 191 G-Sense_Error_Rate -O--CK 100 100 000-48930 Frequent shocks, vibration, unstable base may force heads unloading and following latency in UI. P.S. I have not managed to change heads unload timeout for Toshiba L200 using hdparm. In your case due to high G-Sense_Error_Rate increasing timeout might be dangerous.
Peculiarities of Toshiba L200 2.5in HDD
On 26/10/2022 10:42, David Christensen wrote: P.S. I have not managed to change heads unload timeout for Toshiba L200 using hdparm. In your case due to high G-Sense_Error_Rate increasing timeout might be dangerous. So, you own one of these drives? I have a Toshiba L200 HDWK105 SATA 3 500G CMR disk as a secondary storage. I do not like that heads unload timeout is just a few seconds and such state is hidden, so udisks2 wakes it up every 10 minutes to poll for its SMART state. I see a lot of warnings in docs for HDD tools and I am unsure how to stay on the safe margin during experiments. I do not expect anything directly related to the original subject in this thread, but to not hijack this thread, you may ask your questions e.g. in response to https://lists.debian.org/msgid-search/?m=tbohk8$60u$1...@ciao.gmane.io
Re: Peculiarities of Toshiba L200 2.5in HDD
On 26/10/2022 17:21, Dan Ritter wrote: Max Nikulin wrote: On 26/10/2022 10:42, David Christensen wrote: So, you own one of these drives? I do not expect anything directly related to the original subject in this thread, but to not hijack this thread, you may ask your questions e.g. in response to No, no, you have found the most likely problem. I expect that David may have some questions concerning this disk series. Using a shingled drive with a short unload timeout is definitely going to cause poor performance. I do not think that the original problem is related to partitions. I would blame frequent shock events (and following timeout before allowing heads loading) rather than short unload timeout. Almost 15 years ago I had to spend enough time setting up laptop mode, configuring syslog, etc. when I noticed rapidly increasing heads load count quickly exhausting disk resource. I set some spin down timeout. I can not say that experience was terrible. Perhaps I just expected some freezes and sound of spinning up disk was an indicator that a few seconds later everything would work again. Heads loading is faster, unless blocked due to a recent shock. Side note: It seems disk firmware works great since the disk is not dead after almost 5 hits. I am surprised that such count does not affect the disk wearing (as it is perceived by firmware) at all: notice 100 191 G-Sense_Error_Rate -O--CK 100 100 000-48930I am not going to perform the following experiment with my shiny new disk even though its result may be interesting. Smoothly wave the laptop up and down actively enough to achieve more G-Sense errors (likely accompanied by clanking sound due to heads unloading) and measure read speed during this procedure. An SSD should not have any problem with shocks of vibration having reasonable magnitude. Real reason of poor performance may be particular applications: - some memory hungry one (a browser with a lot of tabs and heavy pages) that causes kernel to drop disk cache, - some program writing to the disk actively enough to overflow 128M buffer of the SMR disk.
Re: Some of the parameters used in my genisoimage command don't produce a bootable ISO image
On 28/10/2022 07:07, Mario Marietto wrote: find . | cpio --create I rarely use cpio, but recently there was a thread on tar and unwanted hard links in the created archive. "find" output mixes regular files and directories. If the archiver recursively walks through received directories then result may differ from expectations.
Re: How to can make a partition in my hard disk ?
On 28/10/2022 05:19, William Torrez Corea wrote: Only use my laptop in my desktop and when i going to sleep in a night table for watch a movie. Compare G-Sense_Error_Rate attribute value in smartctl output in the beginning and at the end. In some buildings a train line nearby may cause strong enough vibration (I do not believe in such hypothesis). On 20/10/2022 08:26, William Torrez Corea wrote: when I start firefox the page stays loading too much time and when I try to open LibreOffice or Gimp it loads the program very slow. In the browser I have different tabs open. Output of the "free" utility may give some clue on memory usage. Start "journalctl -f" before launching application to check for new errors. Watch top lines in "top" statistics (processor and memory "M") while application are starting to see what is the bottleneck: CPU, input/output, or RAM. Do the same for "iotop" to see disk read speed.
Re: tbird broken
On 18/11/2022 14:42, Kamil Jońca wrote:
user_pref("mailnews.nntp.jsmodule", false);
might help
Notice that Thunderbird developers are closing bugs related to the old
NNTP implementation marking them WONTFIX.
Re: tbird broken
On 19/11/2022 18:59, Kamil Jońca wrote:
On 18/11/2022 14:42, Kamil Jońca wrote:
user_pref("mailnews.nntp.jsmodule", false);
...
But old implementation works, at least for me, while new not.
I would say that both implementations have issues and I faced more
painful bugs in the older one.
For example:
https://bugzilla.mozilla.org/show_bug.cgi?id=1788102
Earlier you cited
https://bugzilla.mozilla.org/show_bug.cgi?id=1777439
"new JS NNTP implementation broken in TB 102: news articles are not
shown in the preview pane nor message view tab."
that is related to NNTP servers requiring authentication and closed as
fixed in Thunderbird-102.
Concerning the bug 1788102 "huge articles unreadable on first reading
but show correctly later (once download fully completed?)", I am afraid,
you are out of luck, unless you will manage to provide clear steps to
reproduce. I will not be surprised if the developers decide to drop old
implementation before some later extended support release.
Have you tried to sort messages by size in public newsgroups to find
large messages? Would it "help" to limit network bandwidth using
iptables/nftables? Even setting "mailnews.nntp.loglevel" to "All" and
restarting Thunderbird might shed some light on the issue.
Re: TBird mail (was: user perms)
On 07/12/2022 06:58, gene heskett wrote:
profiledir:
what or where is this "profiledir:"? after sudo updatedb today, I find
Querying search engine "thunderbird profile directory" gives
https://support.mozilla.org/en-US/kb/profiles-where-thunderbird-stores-user-data
http://kb.mozillazine.org/Profile_folder_-_Thunderbird
How to find your profile
- Click on the menu button or menu bar.
- From the Help menu, click Troubleshooting Information.
- In the Application Basics section, Profile Directory, click on Open
Directory.
- The Files window will show the name of the profile as well as the path
to it.
user_pref("mail.quoteasblock", false);
user_pref("mail.quoted_graphical", false);
You may bypass looking up for the profile directory and change or create
any options in Hamburger menu / Settings / Config Editor (at the bottom
of the page).
Re: TBird mail (was: user perms)
On 07/12/2022 15:02, gene heskett wrote: And when I find it. the prefs.js file does not contain any of the strings mentioned in the reply I quoted and everyone but you have clipped. Am I to add these below? According to the top of the file, t-bird must be stopped before editing as it saves this anew when stopping. That is why I suggested you to use built-in config editor. Doesn't your preferred search engine gives relevant result for "thunderbird config editor"? https://support.mozilla.org/en-US/kb/config-editor You may bypass looking up for the profile directory and change or create any options in Hamburger menu / Settings / Config Editor (at ---^^ the bottom of the page).
Re: hard disk and sub flash disk issues
On 08/12/2022 20:38, Semih Ozlem wrote: (i) I accidentally turned one of the partitions on the hard disk of the machine I currently have to linux swap. That partition contained files that I may need to review or use later. Is there a way to recove those files and if so how A chance is rather low. 1. swapoff for this partition (immediately) 2. change partition type to avoid automount it by e.g. live systems, edit /etc/fstab. 3. create an image for this partition. It will allow to restore current state after unsuccessful attempt to restore files. 4. install testdisk and run photorec as root. The tool is able to restore some types of files if they are not fragmented. File names would not be preserved, so high enough efforts may be required to sort recovered files. Unsure if e.g. backup of file system superblock will help to recover at least a part of directory structures. Check actual type of the partition (I am unsure concerning precise action for "turned one of the partitions ... to linux swap") file -s /dev/sdX
bash read and field separator
On 07/12/2022 19:18, Greg Wooledge wrote:
rlart() {
local day time path
find "${1:-.}" -type f -printf '%T@ %TY-%Tm-%Td %TT %p\0' |
sort -zn |
while read -rd '' _ day time path; do
printf '%s %s %s\n' "$day" "${time%.*}" "$path"
done
}
I was not aware of the "read -d ''" feature that allows to work with
null-terminated records, so thank you for the hint. However the read
command strips leading and trailing spaces and newlines from path. What
is the best way to preserve them? I have tried slash (a character
disallowed in file names) instead of space as field separator:
printf 'a///1/1\n\000b/ 2/\000c/3/\000d/4//\000' |
while IFS='/' read -r -d '' num path ; do
printf '%s %q\n' "$num" "$path" ;
done
a $'//1/1\n'
b \ 2
c 3
d 4//
No problem with spaces, leading "//" are preserved, but single trailing
slash is removed after "3".
Another question is if "$@" instead of "${1:-.}" may cause any problem.
Re: bash read and field separator
On 08/12/2022 23:52, Greg Wooledge wrote:
On Thu, Dec 08, 2022 at 09:53:37PM +0700, Max Nikulin wrote:
I have tried slash (a character disallowed in file
names) instead of space as field separator:
The "problem" with that is we have slashes in our pathnames (%p) in
almost all cases. But if we carefully build the read command, I
think it should keep them intact.
I am trying to figure out if it is possible to avoid stripping of single
(and only single) trailing slash. It is not an issue while "-type f"
selector is used, but I am trying to evaluate if it may be used as a
more general approach.
find dir dir/ dir// | sed -n l
dir$
dir/$
dir//$
find dir dir/ dir// -printf '%T@/%TY-%Tm-%Td/%TT/%p\0' | sort -zn |
while IFS=/ read -rd '' _ day time path; do
printf '%s %s %s\n' "$day" "${time%.*}" "$path";
done | sed -n l
2022-12-09 22:13:14 dir$
2022-12-09 22:13:14 dir$
2022-12-09 22:13:14 dir//$
Single trailing / disappeared. Using any other character may cause
incorrect path in the output.
Re: bash read and field separator
On 09/12/2022 23:20, Greg Wooledge wrote:
On Fri, Dec 09, 2022 at 10:23:29PM +0700, Max Nikulin wrote:
find dir dir/ dir// -printf '%T@/%TY-%Tm-%Td/%TT/%p\0' | sort -zn |
while IFS=/ read -rd '' _ day time path; do
printf '%s %s %s\n' "$day" "${time%.*}" "$path";
done | sed -n l
2022-12-09 22:13:14 dir$
2022-12-09 22:13:14 dir$
2022-12-09 22:13:14 dir//$
Single trailing / disappeared. Using any other character may cause incorrect
path in the output.
I can't reproduce your result.
unicorn:~$ mkdir tmp/leaf
unicorn:~$ find tmp/leaf tmp/leaf/ tmp/leaf// -printf '%T@/%TY-%Tm-%Td/%TT/%p\0' | while
IFS=/ read -rd '' _ day time path; do printf '%s\n' "$path"; done
tmp/leaf
tmp/leaf/
tmp/leaf//
It took some time for me to realize why you case is different. Please, try
cd tmp
find leaf leaf/ leaf// -printf '%T@/%TY-%Tm-%Td/%TT/%p\0' | while IFS=/
read -rd '' _ day time path; do printf '%s\n' "$path"; done
I have found some speculations that it is standard behavior
https://stackoverflow.com/questions/52762210/shell-read-sometimes-strips-trailing-delimiter
Unfortunately I can not make similar conclusion after reading the cited
documents. From my point of view
set -- "$var"
is not the same as read. In the latter case remaining part of the line
is assigned to the last variable. The argument that with preserved
trailing separator it is impossible to express single field, does not
work for read as well.
In dash I see the same behavior after dropping unsupported -d and using
\n as record separator.
Re: installation partition recommendations
On 11/12/2022 22:01, Semih Ozlem wrote: Basically the menu that offers choices for where to boot the machine appears. If Debian or USB drive is chosen the menu comes back with no progress at all. Directory structure for UEFI boot (sdd4 in your case) depends on whether it is internal disk or removable storage: EFI/debian/shimx64.efi and EFI/BOOT/bootx64.efi (the same file content). Particular implementation may have bugs. I am unsure which way firmware interprets a USB drive with multiple partitions in your case. If installer created just one directory then I would try to copy it to another name. Do you have an option to choose boot from particular .efi file? If so try bootx64.efi or shimx64.efi directly Within the bios secure boot is disabled. ... or grubx64.efi in this case. Please, show the list of files on sdd4 find EFI | sort If you can boot to some linux on that machine (e.g. a USB drive with live image) then show output of efibootmgr -v
Re: Independent menu system
On 14/12/2022 06:22, pa...@quillandmouse.com wrote: I prefer to run i3wm, but it has no native menu system. Like Openbox, GNOME, Plasma, etc. Does anyone know of a menu system/program which reads *.desktop files, and can supply categorized menus, but doesn't insist on being run under a non-i3wm desktop environment? See the "menu" package. It is debian menu and /etc/menu-methods/fluxbox is an example how window managers may use it.
Re: loss of screen resolution, part 2
On 20/12/2022 09:49, Timothy M Butterworth wrote: Newer Intel graphics require closed source binary blobs. Try installing firmware-linux-nonfree. In the previous thread somebody spotted an issue with fetching modes supported by the monitor. Examples of commands to debug such problem: get-edid | parse-edid edid-decode /sys/class/drm/card0-HDMI-A-1/edid
Re: loss of screen resolution, part 2
On 21/12/2022 00:26, Kleene, Steven (kleenesj) wrote: On Tuesday, December 20, 2022 10:36 AM, Max Nikulin wrote: get-edid | parse-edid edid-decode /sys/class/drm/card0-HDMI-A-1/edid Thanks. get-edid doesn't find any EDIDs, and there are no edid files under /sys/devices. Have you checked "journalctl -b" (current boot only) output for messages related to missing firmware? It may be noticeable in "apt upgrade" messages when initramfs is created during installing of new kernel. As an experiment (it is better to remove new file or restore old version) latest firmware files may be taken from https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree Such changes usually requires "update-initramfs -u" (normally performed by scripts insides firmware .deb packages).
Re: colorscheme in vi
On 23/12/2022 18:36, Pierre Frenkiel wrote: I wanted to open vi with a white background and a black fg to do that, Do you mean graphical gvim window or vim running in a terminal? In the latter case which one and should vim follow terminal background? Is tmux or screen involved? Besides colorscheme there is another setting :set background=dark or :set background=light I do not remember if proper value is reliably chosen during startup e.g. for terminal applications setting TERM=xterm-256color Finally, if colorscheme white works despite the warning, what happens if you commented out or removing this line?
Re: colorscheme in vi
On 24/12/2022 03:49, Pierre Frenkiel wrote: if I put in .vimrc "colorscheme white" or any unknown color (qwert for example) I get a message "cannot find colorscheme xxx", but I have black on white characters with vim. I am curious what is the output of :colorscheme (without argument) when you do not set colorscheme in init files. Is it "default"? to get rid of the message, I replaced the vi binary by a script where stderr is redirected to /dev/null I would say it is quite rude. You may miss a message notifying you that you have some typo in a configuration file or even some severe runtime error.
Re: problem with loading installer (hard disk method)
On 29/12/2022 08:10, lou wrote: set root='(hd0,msdos1)' i boot installer from buster, installed at sda4, i wonder if i shall change hd0 to hd3 in line 4 above If you have multiple hard drives you may need to change hd0, but at first it is necessary to change partition from msdos1 to msdos4 or gpt4. It depends on partition types (label): DOS MBR or GPT. Grub command line allows to try various variants interactively and to boot installer without modifying of the config file.
Re: What layout should I use to emulate a German keyboard but also be able to type other European characters?
On 29/12/2022 18:23, to...@tuxteam.de wrote: I mapped the compose key to CAPS LOCK (who needs that, right?) /etc/default/keyboard: XKBOPTIONS="grp:shift_caps_switch,compose:ralt" - [CapsLock] to select first layout - [Shift+CapsLock] to select second layout - [RightAlt] for Compose Gnome can not handle CapsLock as layout switch though. So 2 or 3 keyboard layouts and Compose for other languages may be an option.
Re: problem with loading installer (hard disk method)
On 29/12/2022 10:55, lou wrote:
i have solved my 2nd problem on my own, thanks anyway!
Great
i have 2 suggestions:
https://www.debian.org/releases/bullseye/amd64/apes02.en.html
E.2. Contributing to This Document
If you have problems or suggestions regarding this document, you should
probably submit them as a bug report against the package
installation-guide. See the reportbug package or read the online
documentation of the Debian Bug Tracking System.
Suggestions sent to this list may be lost unnoticed by the developers.
I suppose, loading installer from hard drive is for advanced users.
There is should be balance for amount of details in Debian manuals. It
is grub manual that should cover the topic thoroughly.
Notice that you having an .iso file stored on a hard drive, it is
possible to boot it directly. A couple of years ago I tried
menuentry "ISO Ubuntu-20.10 amd64" {
set isofile="/ubuntu/ubuntu-20.10-desktop-amd64.iso"
loopback loop (hd0,5)$isofile
linux (loop)/casper/vmlinuz boot=casper
iso-scan/filename=$isofile noprompt noeject toram
initrd (loop)/casper/initrd
}
linux and initrd commands are taken from the grub configuration inside
the iso file.
The grml-rescueboot package uses the same approach.
Re: problem with loading installer (hard disk method)
On 30/12/2022 00:42, Brian wrote: I doubt GRUB would act on (hd0,partition1). Perhaps the idea was that "partition1" may be "msdos1" or "gpt1". I do not have some specific suggestion how to improve https://d-i.debian.org/manual/en.amd64/ch05s01.html#boot-initrd On Thu 29 Dec 2022 at 22:07:15 +0700, Max Nikulin wrote: Notice that you having an .iso file stored on a hard drive, it is possible to boot it directly. ... GRUB's loopback facility will not work with installer images. They do not contain iso-scan. https://bugs.debian.org/618000 https://bugs.debian.org/724931 My bad, thank you for clarification. Is there a reason why these report are not merged? I rarely need a "hardware" install media, for a VM an .iso file can be used directly to emulate cdrom. I believed that Ubuntu uses features available everywhere. Actually when I need a bootable USB, I use non-destructive method: remove old files and extract content of an .iso file using 7zip or bsdtar. It is enough for UEFI, for BIOS a couple of additional steps were required to install syslinux and to adjust its config.
Re: Why Debian packaging structure is so difficult
On 09/01/2023 13:20, Sadhu Santh wrote: If not, why the upstream structure is not made simpler? If I remember correctly, current repository layout minimizes traffic at the moment of new release. .deb packages are already synced as a part of testing suite. Almost certainly it is explained on some official page or in a manual, but I can not figure out where I saw it. For tools aware of Debian repository structure see https://www.debian.org/mirror/ftpmirror https://wiki.debian.org/DebianRepository/Setup#Debian_Repository_Mirroring_Tools Personally I use apt-cacher-ng, so I can not comment reliability of other tools.
Re: Setting up bindfs mount in LXC container
On 17/01/2023 04:06, Richard Hector wrote:
I'm using bindfs in my web LXC containers to allow particular users to
write to their site docroot as the correct user.
I am not familiar with bindfs, so I may miss something important for
your use case.
First of all I am unsure why you prefer bindfs instead of mapping some
container users to host users using namespaces. With the following
configuration 1000 inside a container and on the host is the same UID:
lxc.idmap = u 0 10 1000
lxc.idmap = u 1000 1000 1
lxc.idmap = u 1001 101001 64535
lxc.idmap = g 0 10 1000
lxc.idmap = g 1000 1000 1
lxc.idmap = g 1001 101001 64535
lxc.mount.entry = /home/richard/sitename/doc_root /srv/sitename/doc_root
none bind,optional,create=dir
In /usr/local/bin/fuse.hook:
I would look into lxcfs hook for inspiration
In /usr/local/bin/fuse.hook.s2:
lxc-device -n ${LXC_NAME} add /dev/fuse
Is there any reason why it can not be done using lxc.mount.entry in the
container config?
lxc-attach -n ${LXC_NAME} /usr/local/bin/bindfs_mount
I would consider adding a systemd unit inside container. Unsure if could
be done using an udev rule.
If I don't use the at job, but run those commands manually after boot,
it works fine with no error messages.
Unsure if it is relevant, but it is better to run lxc-start and
lxc-attach as a systemd unit with Delegate=yes configuration, either a
temporary one (systemd-run) or configured as a service. It ensures
proper cgroup and scope. Otherwise some cryptic errors may happen.
Re: Setting up bindfs mount in LXC container
On 17/01/2023 17:52, Max Nikulin wrote: lxc.mount.entry = /home/richard/sitename/doc_root /srv/sitename/doc_root none bind,optional,create=dir Sorry, path inside the container should be without the leading slash. lxc.mount.entry = /home/richard/sitename/doc_root srv/sitename/doc_root none bind,optional,create=dir
Re: Setting up bindfs mount in LXC container
On 18/01/2023 03:52, Richard Hector wrote:
On 17/01/23 23:52, Max Nikulin wrote:
lxc.idmap = u 0 10 1000
lxc.idmap = u 1000 1000 1
lxc.mount.entry = /home/richard/sitename/doc_root
srv/sitename/doc_root none bind,optional,create=dir
My goal is not to map container users to host users, but to allow a
container user (human user) to access a directory as another container
user (non-human owner of files). This should also be doable for multiple
human users for the same site.
Do you mean mapping several users (human and service ones) from a single
container to the same host UID? The approach I suggested works for 1:1
mapping. Another technique is group permissions and ACLs, but I would
not call it straightforward. A user may create a file that belongs to
wrong group or inaccessible by another user.
I would look into lxcfs hook for inspiration
Interesting; will do. Not sure exactly where to start, but will get there.
/usr/share/lxcfs/lxc.mount.hook (lxcfs package)
lxc-device -n ${LXC_NAME} add /dev/fuse
Is there any reason why it can not be done using lxc.mount.entry in
the container config?
Is that usable for adding a device file? The only way I found to do that
is using lxc-device from outside the container. mknod inside doesn't work.
I use it to pass /dev/dri and /dev/video0 to a container. The only issue
is that the webcam is a USB device, so permissions become wrong after
suspend to RAM and the following resume.
lxc-attach -n ${LXC_NAME} /usr/local/bin/bindfs_mount
I would consider adding a systemd unit inside container. Unsure if
could be done using an udev rule.
That might be better, but it does need to rely on the device existing
first.
I see no difference since in your approach device presence is ensured by
a hook. With a lxc.mount.entry it should be even more reliable.
So even for running stuff manually, run it from systemd? Interesting,
will investigate further. I wasn't aware of systemd-run.
I use unprivileged containers to experiment with applications packaged
for other linux distributions or just in relatively isolated
environment. My experience that even after searching for particular
errors it is not obvious what should be done to fix it.
Delegate=yes is mandatory for lxc-start. To attach:
systemd-run --user --scope --property=Delegate=yes -- \
lxc-attach --clear-env --keep-var TERM ...
At first I tried --pty instead of --scope for interactive sessions, but
I noticed some issues as well.
-
https://discuss.linuxcontainers.org/t/pam-cgfs-how-to-setup-with-cgroup-v2/10741/
"Pam_cgfs How to setup with cgroup v2?"
- https://github.com/lxc/lxc/issues/3221#issuecomment-648829685
"Unable to start an unprivileged container on fresh install of Fedora
31 · Issue #3221 · lxc/lxc"
Re: Setting up bindfs mount in LXC container
On 18/01/2023 13:02, Richard Hector wrote:
I have a Wordpress site. The directory /srv/sitename/doc_root, and most
of the directories under it, are owned by user 'sitename'.
PHP runs as 'sitename-run', which has access (via group 'sitename') to
read all of that, but not write it. Some subdirectories, eg
.../doc_root/wp-content/uploads, are group-writeable so that it can save
things there.
An authorised site maintainer, eg me ('richard') (but there may be any
number of others), needs to be able to write under /srv/sitename,
Just an idea (likely should be applied recursively)
setfacl -m user:richard:rwx /srv/sitename
setfacl -m default:user:richard:rwx /srv/sitename
the same for PHP read-only access. However likely it is impossible to
set default: x for directories only. Another issue is that a directory
moved from outside will not inherit default ACL. So despite it looks
like standard problem that may be solved by ACLs I do not have a
complete recipe.
Are there users that have access to container only using ssh? If you
have access to host, you can run any command to access filesystem
outside of container using lxc-usernsexec (perhaps with modified uidmap)
or inside container using lxc-attach ... su --login ...
If have not read it yet, some general LXC ideas are described in
https://stgraber.org/2013/12/20/lxc-1-0-blog-post-series/
Re: Postfix fails after upgrade from bullseye to bookworm
On 21/01/2023 01:55, Charles Curley wrote: root@white:~# ps aux | grep -i openssl root 4586 5.8 0.9 8256 2064 pts/3S+ 11:48 0:00 grep --colour=auto -i openssl root 4587 150 2.1 4720 ?R11:48 0:00 /usr/bin/openssl x509 -subject_hash_old -fingerprint -noout -in QuoVadis_Root_CA_2.pem ...> I have no idea what that's about. Maybe someone with SSL experience can chime in here? Delays related to cryptography may be caused by lack of entropy. You may try to check cat /proc/sys/kernel/random/entropy_avail and ls -l /proc//fs when such delay happens. Some network activity may increase entropy enough to unblock reading random bytes. If my guess is correct then the issue may be alleviated by installing the haveged package. VMs are affected more than hosts.
Re: Unexpected permission denied
On 25/01/2023 21:52, Greg Wooledge wrote:
On Wed, Jan 25, 2023 at 03:36:33PM +0100, Yassine Chaouche wrote:
runas_wwwdata ()
{
echo su - www-data -s /bin/bash -c "$*";
su - www-data -s /bin/bash -c "$*"
}
...
su(1) is pretty much the WORST possible choice for this, as it forces
you to launch a shell, instead of just executing a command directly.
...
unicorn:~# runas() { local who=$1; shift; su - "$who" -s /bin/bash -c "$*"; }
unicorn:~# runas www-data ls -ld /tmp/'file with spaces'
su: warning: cannot change directory to /var/www: No such file or directory
Greg, I agree with your warnings. Just out of curiosity, is there a
reason why the following variant may still be unsafe?
runas() { local who=$1; shift; su --login "$who" --shell=/bin/bash
--command='"$0" "$@"' -- "$@"; }
Re: Unexpected permission denied
On 26/01/2023 11:04, Greg Wooledge wrote:
On Thu, Jan 26, 2023 at 10:26:34AM +0700, Max Nikulin wrote:
Greg, I agree with your warnings. Just out of curiosity, is there a reason
why the following variant may still be unsafe?
runas() { local who=$1; shift; su --login "$who" --shell=/bin/bash
--command='"$0" "$@"' -- "$@"; }
1) http://jdebp.info/FGA/dont-abuse-su-for-dropping-privileges.html
Greg, thank you for the link. Actually I use su namely to create PAM
session, e.g. to get shell inside a container using "lxc-attach -n
container ... -- su - user" (when where is no point to use ssh). setpriv
is likely not an option in such case. Mostly I use sudo. Sometimes I use
the trick with positional arguments passed to "sh -c", but I have never
combined it with su before.
3) --command='"$0" "$@"' is ... very unintuitive, even for an experienced
shell user.
...
su "$who" -c '"$@"' -- x "$@"
To get meaningful command name in ps output, perhaps it is better to use
something like
su "$who" -c '"$@"' -- su-bash "$@"
(or "su-$SHELL" that skill may not be precise). Repeating $1 is likely
worse, since the process is shell, not the passed command.
su "$who" -c '"$@"' -- "$1" "$@"
Can't say
I've ever seen su contortions like this before.
I have an idea what should be referred to as a real abuse of su. Do not
do it, it is just a joke. However it allows to avoid issues with spaces
and other shell specials in arguments.
runas() { local who=$1; exe="$(type -P "$2")"; shift 2; su - "$who" -s
"$exe" -- "$@"; }
Re: Syslog/Rsyslog/Systemctl issue
On 31/01/2023 16:57, deb...@sioban.net wrote: I'm contacting you because I'm clueless on what's happening. Basically my issue is I don't have logs anymore :/ ... > systemctl start rsyslog.service A dependency job for rsyslog.service failed. See 'journalctl -xe' for details. ❯systemctl --failed UNIT LOAD ACTIVE SUB DESCRIPTION ... ● systemd-journald.socket loaded failed failed Journal Socket This line looks suspicious. I would ensure at first that journald is running correctly. Rsyslog likely receives logs from journald. Are logs available through the following commands (last entries and for current boot)? journalctl -e journalctl -b Does systemctl status systemd-journald.service reports any errors?
Re: ASCII formatting for plain text email
On 01/02/2023 01:49, Pierre Willaime wrote: I would like to format plain text emails to increase readability and information separation. The following message is result of ascii export from Emacs Org mode, the source file is attached. Export backend is customizable. https://list.orgmode.org/87bktn8f3y@gnu.org/ [ANN] EmacsConf 2022 Call for Participation. Sun, 17 Jul 2022 20:00:01 -0400
Re: Web page management. Was: Re: Debian release criteria.
On 30/01/2023 10:32, Celejar wrote: For those who may not be aware, the developer of uBlock Origin and uMatrix, Raymond Hill, has ceased work on uMatrix: https://github.com/uBlockOrigin/uMatrix-issues/issues/291#issuecomment-694988696 It is really sour. On a lot of sites, including github, high CPU usage in the case of disabled JavaScript caused by loading animations: CSS or gifs.
Re: Syslog/Rsyslog/Systemctl issue
On 01/02/2023 21:48, Freyja wrote: Instead of just creating the folders, I've reinstalled some packages: I would consider complete reinstall (install from scratch). It may be faster even for a person having experience with systemd troubleshooting. Something low level is affected. However I find the problem interesting: how to get logs if journald can not start.
Re: kernel errors
On 03/02/2023 01:47, Richmond wrote: It might be a good way for someone to reproduce the error on some other machine. I have no problems with the CD/DVD writer and have used it a few times recently. Do you see the same errors if kernel command line is edited from grub to pass non-existing UUID specified in the resume=UUID=... argument? It might be a quick way to reproduce the issue. Is it realistic that usually optical drives are skipped during UUID searches, but specific driver exposes a wrong property, so the device is considered as a regular disk drive?
Re: OT: repair/replace cell in Li-ion battery?
On 05/02/2023 17:37, Yvan Masson wrote: I don't know anything about battery, but I can do easy soldering. 1. Proper technology for rechargeable batteries is welding, not soldering (to prevent overheating). 2. If you just disconnect a cell then a chip monitoring health status of the assembly will detect failure and will prevent further charging (to avoid risk of fire). Sorry, that is all I know. If you are still going to do it yourself, try to find repair guides.
Re: kernel errors
On 05/02/2023 03:12, Richmond wrote: The errors about sr0 come before the stuff about resume. Does the following command generate similar errors (taken from initrd scripts, UUID is intentionally not from the set of existing partitions)? blkid -l -t UUID=---- -o device I would try it at first in normally running system. Perhaps some caches may be involved preventing query to the optical drive. I have no bright idea how to cause drop to initrd shell to try the command there, perhaps root=... kernel command line parameter may be removed or changed to non-existing device from grub prompt. I am unsure which particular command from initrd scans devices, my guess may be wrong.
Re: OT: repair/replace cell in Li-ion battery?
On 05/02/2023 23:32, gene heskett wrote: Speaking as a CET with some knowledge of battery degradation, the thought of replacing an individual cell in a li-ion battery scares me. That new battery will have a bit more capacity, which in turn may force other older cells into reverse polarity, accelerating the rate of failure of the rest of the cells, with the possibility of starting a fire you can't put out. IOW, don't even consider it. Battery rebuilders with any care for the result, may and should select the cells with a 1 % or so like capacity, meaning your rebuilt battery s/b safe, but may not have the labeled capacity. Does your charger know about the reduced capacity? I would not be surprised if nowadays a laptop battery (at least original one) has a balancing circuit that compensates difference of cells capacity e.g. due to non-uniform degradation. By the way, Ottavio may be right. Some laptops have UEFI settings to maintain battery charge level of approximately 70%. It may mitigate degradation if laptop is mostly working from AC adapter.
Re: No USB with qemu+macOS+USB+iPads
On 02/02/2023 23:00, stand...@gmx.net wrote: (qemu) info usbhost Bus 2, Addr 4, Port 9, Speed 480 Mb/s Class 00: USB device 058f:6387, Mass Storage Bus 2, Addr 3, Port 3, Speed 1.5 Mb/s Class 00: USB device 046d:c34b, USB Keyboard Bus 2, Addr 2, Port 2, Speed 12 Mb/s Class 00: USB device 046d:c52b, USB Receiver Does the user running VM have read/write access to the usb device? The docs mention chown, I used sudo -A setfacl -m u:`id -un`:rw /dev/bus/usb/002/007 and in qemu monitor device_add usb-host,hostbus=2,hostport=7,id=usbcam However it was with linux guest.
Re: kernel errors
On 06/02/2023 00:11, Richmond wrote: Max Nikulin writes: On 05/02/2023 03:12, Richmond wrote: The errors about sr0 come before the stuff about resume. Does the following command generate similar errors (taken from initrd scripts, UUID is intentionally not from the set of existing partitions)? blkid -l -t UUID=---- -o device It didn't produce any output. Sorry, I have no motivation to dig into /usr/share/initramfs-tools deeper. Perhaps you may try debug kernel option in combination with bad UUID for resume from grub ... debug resume=UUID=---- with hope to get more verbose output around of optical drive access. You may try to report a bug against initramfs-tools that errors messages are unhelpful and even confusing when resume configuration is inconsistent due to a mistake. Perhaps resume argument may be validated during generation of initrd, however in some cases it will give false positive.
Re: No USB with qemu+macOS+USB+iPads
On 08/02/2023 01:15, stand...@gmx.net wrote: Max Nikulin schrieb am Montag, 6. Februar 2023 um 13:30:06 UTC+1: sudo -A setfacl -m u:`id -un`:rw /dev/bus/usb/002/007 I think running as root doesn't requirte this change, or? Then I have no idea. Does it work if you pass a USB device to a Linux guest (e.g. boot some live image)?
Re: No USB with qemu+macOS+USB+iPads
On 09/02/2023 01:23, Charles Curley wrote: On Wed, 8 Feb 2023 22:07:37 +0700 Max Nikulin wrote: Does it work if you pass a USB device to a Linux guest (e.g. boot some live image)? Yes. I can hand a USB memory stick to a Debian guest, Debian installer, bullseye and bookworm. Host is bullseye. One must still mount it on the guest. My question was addressed to stand...@gmx.net with hope to determine if the problem is related to qemu configuration or to missing drivers in macOSX. Unsure if qemu allows to change the kind of USB bridges (similar to e.g. ethernet adapters). Concerning USB configuration, devices are specified in different ways. stand...@gmx.net: -device usb-host,vendorid=0x8086,productid=0x0808 Charles Curley: -device usb-host,hostdevice=/dev/bus/usb/002/006,id=hostdev0,bus=usb.0,port=4
Re: New Dell Inspiron 15 3000 Series 3511 sound problem on Debian 11 Stable
On 12/02/2023 05:46, Default User wrote: [14777.998904] Modules linked in: ctr ccm nls_ascii nls_cp437 vfat fat ext4 mbcache jbd2 ses enclosure scsi_transport_sas rfcomm cmac algif_hash algif_skcipher af_alg bnep x86_pkg_temp_thermal intel_powerclamp coretemp rtw88_8821ce btusb kvm_intel rtw88_8821c btrtl rtw88_pci btbcm rtw88_core btintel bluetooth kvm mac80211 jitterentropy_rng dell_laptop intel_rapl_msr irqbypass mei_hdcp cfg80211 drbg intel_cstate intel_uncore ansi_cprng dell_wmi dell_smbios ecdh_generic efi_pstore iTCO_wdt pcspkr serio_raw dcdbas rfkill intel_pmc_bxt ecc dell_wmi_descriptor sg libarc4 wmi_bmof joydev iTCO_vendor_support crc16 watchdog ee1004 hid_multitouch snd_sof_pci snd_sof_intel_byt snd_sof_intel_ipc snd_sof_intel_hda_common snd_sof_xtensa_dsp snd_sof snd_sof_intel_hda snd_hda_codec_hdmi snd_soc_hdac_hda snd_hda_ext_core snd_hda_codec_generic snd_soc_acpi_intel_match snd_soc_acpi ledtrig_audio snd_hda_intel snd_intel_dspcfg soundwire_intel soc_button_array int3403_thermal soundwire_generic_allocation I may be wrong, but this line may be a sign of kernel crash causing arbitrary following failures. I would inspect output of journalctl around "Modules linked in" lines. Perhaps --boot=-1 (and -2, -3, etc.) may be added to filter logs for previous boots.
Upgrades of Debian and Ubuntu
On 16/02/2023 14:59, Nicolas George wrote: Debian is not Ubuntu, major upgrade do not break the system. Perhaps the issue is not Ubuntu per se, but e.g. number of 3rd-party repositories. Ubuntu users may be more frequently affected by suggestions to add a ppa repository. I have heard opinions about broken upgrades of ubuntu, but real reasons are unclear for me. My experience is that without 3rd party repository upgrades works well enough. I faced some minor issues, but they were not a reason to reinstall.
Why users uninstall avahi
On 14/02/2023 22:48, David Wright wrote: On Mon 13 Feb 2023 at 11:39:31 (+0100), Christoph Brinkhaus wrote: Today I have deleted almost everything of avahi and xfce4. After a reboot the 169.254.xxx.yyy is no more configured. Yes, and that's a problem for anyone trying to replicate the configuration of these addresses: we usually never see files and logs from offending systems, but just reports of package deletion or, even less helpful, so-called nuking of random files. Are there design overview what avahi-autoipd is supposed to do and which way dhcp clients interact with it? A dozen of years ago I purged avahi packages because they were adding more complexity to debugging of network issues. Avahi and 169.254.x.y addresses are noticed by users when network does not work. It may be caused by various reasons: provider's DHCP server is temporary unavailable, problems during negotiation for connection type and bandwidth, maybe ethernet card driver. Since other part is involved, it may be a problem to reproduce the issue when it is enough time to debug it. That times I did not find info when avahi may be useful for me. Quick search for zeroconf purpose did not result in useful links. Recently I saw a 169.254.x.y address when I decided to take advantage of 1Gb ethernet ports in laptops to copy some large files. RTL8723BE WiFi sometimes may work unreliably. NetworkManager allows to start a DHCP server from UI by creation of shared connection. Unfortunately due to dust on rarely used ethernet connector connection was downgrading from 1Gb to 100Mb and NetworkManager considered it as failure. As a result on the other side avahi-autoipd assigned its address. So to get 169.254.x.y address, ensure physical connection with no DHCP server. In real life it happens when you should complete some other tasks and you are upset due to broken network. From my point of view, to improve reputation of Avahi, it is necessary to clearly explain the role of avahi-autoipd. Maybe better docs have appeared last years, I have not checked.
Re: Virtual machine affects client screen resolution
On 18/02/2023 08:57, Albert S. wrote: If you think I issued the xrandr command on the wrong machine, that was not the case: history makes it clear. Behavior depends on the DISPLAY environment value at the moment when xrandr was executed. Likely it was pointed not to vnc Xserver, but to the socket forwarded to your workstation by ssh. Try echo "$DISPLAY" In a local shell, in a ssh session, and from vncviewer.
Re: Partitioning an SSD?
On 16/02/2023 22:25, Joe wrote: Stretch installed perfectly dual-boot with Win 10 on an EFI Acer netbook, but upgrading to Buster broke booting to grub. It actually broke EFI booting completely, but I've been able to restore booting at least to Windows. And yes, I've tried everything the Net can suggest. UEFI implementation by particular vendor may be buggy. E.g. an old HP laptop may be booted up either when it is configured to load a specific .efi file or when directory structures on EFI internal hard drive partition follows layout designed for a removable storage. It ignores most of EFI variables, so e.g. it is better to remove fbx.efi fallback intended to fix boot order. When you have learned such stuff, it is easy to fix boot. Unfortunately pieces of the puzzle was spread over multiple resources. I am not surprised that installer was unable to apply proper workarounds on particular machine. I have heard that vendors may add code intended to ensure booting Windows, but other OSes may be out of luck using the same file layout.
Re: snapd vs apt
On 17/02/2023 00:48, winnie hw wrote: When both snapd and apt sources are available, which one should I choose for package installation? Though I found the package versions in snapd are a bit newer than apt. With snapd you do not have full control when an application is updated. The daemon decides it for you. It may matter during debugging of some issue. Attempt to reproduce it in a clean environment may fail because on latest version is available in snap. With apt it is possible to setup a custom mirror, e.g. apt-cacher-ng proxy. A proxy for snap must be registered in Canonical and requires network connection to their repository. When some application is packaged for Linux distributions it is easier to rebuild it with a custom patch. I suspect that a developer distributing a snap package may use specific and not well documented build environment.
Re: hard disk installation method fails
On 18/02/2023 12:26, Charles Curley wrote:
menuentry "Install Debian 12 OS (preseeded)" {
set root='hd0,msdos5'
set isofile=/firmware-testing-i386-netinst.iso
insmod part_msdos
insmod loopback
loopback loop (hd0,msdos5)$isofile
linux (loop)/install.386/vmlinuz auto=true file=/media/preseed.cfg
initrd (loop)/install.386/initrd.gz
}
Due to the following message I am confused if it should work for Debian:
Brian to debian-user. Re: problem with loading installer (hard disk
method) Thu, 29 Dec 2022 17:42:59 +.
https://lists.debian.org/msgid-search/?m=29122022172846.4fc3cfd29...@desktop.copernicus.org.uk
GRUB's loopback facility will not work with installer images.
They do not contain iso-scan.
https://bugs.debian.org/618000
https://bugs.debian.org/724931
Re: snapd vs apt
On 18/02/2023 13:33, tomas wrote: On Sat, Feb 18, 2023 at 12:50:12PM +0700, Max Nikulin wrote: When some application is packaged for Linux distributions it is easier to rebuild it with a custom patch. I suspect that a developer distributing a snap package may use specific and not well documented build environment. I think this one is a very important point which hasn't been addressed yet in this thread: Debian packages provide not only their source, but the whole build setup to rebuild that package (and most of them in a reproducible [1] way, i.e. they yield bit-for-bit identical binaries whenever the sources and the build dependencies stay constant). In my opinion reproducible build is a next order problem. I have seen phrases like "it is too complicated to build". I am in doubts if such application can still be called free software despite its license. It is open source, but it is prohibitively difficult to build. I do not consider maintaining build rules as a pleasant activity during software development. Keeping scripts portable may require enough efforts. Certainly it is great when users do not have to compile applications themselves. Unfortunately convenient channels to deliver ready to use software caused emergence on an opposite issue. Nobody besides its developer can build it. It is expected that for the developer build scripts have lowest priority since there is no backpressure from maintainers of Linux distributions. Alleviation of an issue creates new challenges. More snap issues: I am unsure if it is possible to get debug symbols for snap packages. Attaching debugger to a sandboxed process may be not so easy as well. I have not investigated if I can tune level of isolation of snap packages for specific purposes. E.g. firefox snap does not allow opening PDF files from /usr/share/doc because system directories are isolated. On the other hand whole home directory is exposed for convenience. Can I start a browser instance with specific network namespace with firewall rules blocking access to local resources while another browser profile have access to the local network? Snap provides some knobs, however I an unsure that I will not face hard coded limitations. Not to mention complains that due to compressed image start time may became an issue.
Re: CUPS - how to match autodetected printers to physical ones
On 16/02/2023 22:41, Greg Wooledge wrote: 2) Also suggested: avahi-browse -rt _ipp._tcp This year, the output of that command no longer contains my printer's IP address. Last year, it did. I have no idea why this has changed. Avahi was mentioned in the ipv6 thread, so I decided to read links on the Debian wiki page https://en.wikipedia.org/wiki/Avahi_%28software%29 Avahi's performance resembles that of Bonjour, sometimes exceeding it; however Avahi can lose services when managing large numbers of requests simultaneously.[3] 3. Analysis of Peer-to-Peer Protocols Performance for Establishing a Decentralized Desktop Grid Middleware http://lipn.fr/~cerin/documents/PresentationSGS08heithem.pdf Unsure if it is relevant 4) Also mentioned: port 9100. For grins, I did "telnet 10.76.172.100 9100" and after that connected I typed "HELLO WORLD", then pressed Enter, then Ctrl-] q Enter to close the telnet session. I am curious if there is some tool convenient to get report concerning various service service discovery protocols provided at some ip (mDNS-SD, uPnP, etc.). Queue \\SPS\S010NEURD14841M Does anyone know what technology is used by windows nowadays (Simple Service Discovery Protocol, SSDP?) and which tools can be used in Linux to get list of services?
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 19/02/2023 23:35, Christoph Brinkhaus wrote: Am Sun, Feb 19, 2023 at 05:21:47PM +0100 schrieb Geert Stappers: Having installed package openvswitch-switch and doing `ip route` I do get 169.254.0.0/16 dev ovs-system scope link src 169.254.201.7 metric 1004 Please have a look at https://wiki.debian.org/Avahi. I hope, somebody with more knowledge of related technology will correct me. I find it confusing that the wiki page neither mentions avahi-autoipd nor has a link explaining interaction of avahi and avahi-autoipd. My impression is that the purpose if Avahi is discovery of services in multicast network segment and publishing services available on the host where avahi daemon is running to make them available for other computers. Its scope is .local host names. IP address may be received from centralized DHCP server. 169.254.x.y addresses are link local (IPv4LL) and usually appears when IP address is not configured and an attempt to get it through DHCP fails. Such addresses may be configured by avahi-autoipd, unsure concerning systemd(-networkd?). So to avoid 169.254.x.y addresses, it necessary to disable link local addressed (avahi-autoipd). My guess is that Avahi as service discovery tool may still work when usual (static or DHCP) IP address is configured. Perhaps to get rid of 169.254.x.y addresses, it is enough to properly configure network interface, either to ensure that DHCP server is available or to assign a static address. After that you may forget about existence of avahi-autoipd.
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 20/02/2023 01:57, Geert Stappers wrote: feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: ovs-system: Failed to update link state, ignoring: No such file or directory feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: ovs-system: Failed to update link state, ignoring: No such file or directory feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: ovsbr0: Failed to update link state, ignoring: No such file or directory feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: ovsbr0: Failed to update link state, ignoring: No such file or directory feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: ovsbr0: Failed to update link state, ignoring: No such file or directory feb 19 18:48:17 trancilo systemd-networkd-wait-online[601]: Timeout occurred while waiting for network connectivity. Which way IP address is supposed to be configured for the ovs-system network device? My guess is that there will be no 169.254.x.y address and route when this failure is fixed. I am unsure if it is possible to configure avahi-autoipd (or another zeroconf daemon actually running) to ignore specific nework interface. The following is unrelated to the original question. I am curious if Avahi (not autoipd) supports IPv4LL addresses on multiple network interfaces. Likely its primary use case is a laptop connected to WiFi (or with plugged in ethernet cable), not a router or a host having complex network configuration for the sake of VMs.
Re: Fw: python thinking at debian
On 20/02/2023 20:59, Vic Cekvenich wrote: ... I think that is the goal for all of us. ... I did/do apt install python3-pip and python3-venv and python ensurepip still does not work. But if I do pyenv to build, all of python works well. There is large gap between too generic goal "can develop" and too specific issue that ensurepip does not work for you. Details related to intermediate level task is missed, so it is difficult to say if it is a real problem or just a bit different way for similar actions is assumed by developers. What you consider as restrictions may help to ensure that other packages can rely on sane enough environment. Debian bug tracker or mailing list archive may contain discussion why some limitations were added and what approach (venv, etc.) allows to get work done. /usr/share/doc/python3.*-venv/README.Debian /usr/share/doc/python3/python-policy.txt.gz May contain some hints.
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 20/02/2023 21:44, Christoph Brinkhaus wrote: Am Mon, Feb 20, 2023 at 09:59:20AM +0700 schrieb Max Nikulin: Perhaps to get rid of 169.254.x.y addresses, it is enough to properly configure network interface, either to ensure that DHCP server is available or to assign a static address. After that you may forget about existence of avahi-autoipd. On my system it did not help. One "issue" might be, that systemd starts services in some sequence. But it does not wait for a service to complete. At least in case of stuff I have observed on my system. Out of curiosity, is link-local IP address assigned during boot or later when e.g. WiFi connection is temporary lost? How long does it take to get response from DHCP server? Which way network is configured (ifupdown, NetworkManager, systemd-networkd) in your case?
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 22/02/2023 01:26, Christoph Brinkhaus wrote: I have no idea if it is possible to estimate a DHCP response time. Since static IP address is assigned, it does not matter. I expected DHCP configuration and that delay may be noticed in `journalctl -b 0` logs. [Unit] Description=A remote mail retrieval and forwarding utility After=network-online.target opensmtpd.service unbound.service Requires=opensmtpd.service unbound.service But fetchmail starts before the dependencies have been finished. I can not say that I fully understand interaction of After and Requires/Wants options. I would try additional Wants=network-online.target [Match] Name=w* [Network] DHCP=no Address=192.168.0.62/24 Gateway=192.168.0.32 DNS=127.0.0.1 There are options like RequiredForOnline, see systemd.network(5), but likely default value is yes. However avahi-autoipd should be started concurrently with network configuration to assign link-local address in the case of failure.
Re: Where to look into when system crashes
On 22/02/2023 11:34, Qiming Ye wrote: I have a Debian box running for 280+ days without rebooting, recently it crashes pretty much everyday. Where should I look for the reason of crashing? sudo journalctl You may limit logs to specific boot by adding option like "--boot=-1". Scroll to end [G] or add "--end" flag to see last messages before crash, however it is better to pay some attention to all errors.
Re: Virtual machine affects client screen resolution
On 19/02/2023 01:01, Roy J. Tellason, Sr. wrote: On Saturday 18 February 2023 12:17:20 am Max Nikulin wrote: echo "$DISPLAY" So this got me curious, and I tried it out. In the terminal that's running inside of the virtualbox instance where I'm doing emails, it comes back with: :0 Have you tried to emulate multiple monitors in virtualbox? But in a terminal which is running on the host Debian system, it comes back with: :0.0 I wonder why the difference? My guess is that it may depend on graphics adapter and its driver. I have heard that a display may have several screens (it is not the same as multiple monitors that show different regions of the same display and screen). I have never tried such configuration. I am unsure which approach is used in nvidia's xinerama.
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 22/02/2023 23:45, Christoph Brinkhaus wrote: Am Wed, Feb 22, 2023 at 10:24:59PM +0700 schrieb Max Nikulin: On 22/02/2023 01:26, Christoph Brinkhaus wrote: [Unit] Description=A remote mail retrieval and forwarding utility After=network-online.target opensmtpd.service unbound.service Requires=opensmtpd.service unbound.service ... In case of my fetchmail setup the culprit is unbound. At the startup of unbound it takes some time to exchange keys and so on. I have no experience with unbound and I am not sure at which moment it notifies systemd that the service is ready. However I have found a recent bug https://github.com/NLnetLabs/unbound/issues/773 "When used with systemd-networkd, unbound does not start until systemd-networkd-wait-online.service times out" Perhaps the package in Debian has an older version of the unbound.service file and so is not affected. ... However avahi-autoipd should be started concurrently with network configuration to assign link-local address in the case of failure. In a different thread - it was about IPv6 which has mutated slightly - several users claimed that the avahi-autoip is useful for their business. I mean IPv4 link local addresses 169.254.x.y. My impression is that avahi-autoipd was created for the cases when there is no point to setup centralized DHCP server. On the other hand I agree that a router (and so DHCP out of the box) is more wide spread configuration than connecting a couple of devices directly or through a switch.
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 25/02/2023 04:43, Geert Stappers wrote: Having `apt purge avahi-autoipd` still gets me "auto IPv4 address" Ideas how to avoid it are welcome. Have you checked "journalctl --boot" for logs which component assigns 169.254.x.y address and for various errors related to network? I am not familiar with openvswitch (or another package that really created ovs-system and ovsbr0), so I have no idea how they may be configured (systemd-networkd, NetworkManager, netplan, ifupdown) in your case and how to configure them properly. Perhaps it better to ask their community how to avoid failure of systemd-networkd-wait-online and assigning of IPv4LL address. I have realized that your problem may be more general than just ovs-system, I do not like the following log line and which file is not found is not clear for me: feb 19 18:46:18 trancilo systemd-networkd-wait-online[601]: wlan0: Failed to update link state, ignoring: No such file or directory
Re: Virtual machine affects client screen resolution
On 25/02/2023 00:55, Roy J. Tellason, Sr. wrote: On Wednesday 22 February 2023 09:24:17 pm Max Nikulin wrote: On 19/02/2023 01:01, Roy J. Tellason, Sr. wrote: So this got me curious, and I tried it out. In the terminal that's running inside of the virtualbox instance where I'm doing emails, it comes back with: :0 Have you tried to emulate multiple monitors in virtualbox? I'm not sure what I need to do with the computer to make this happen. VirtualBox can emulate multiple monitors, they may be represented as several windows on the same physical device. It is convenient to test behavior of applications in the configuration with multiple monitors. There is an option in VM configuration UI. But in a terminal which is running on the host Debian system, it comes back with: :0.0 I wonder why the difference? My guess is that it may depend on graphics adapter and its driver. It's an older machine with a VGA output being used. I assume that I'll need to get some kind of a card with an HDMI output and a cable to make that happen. No idea what the driver is, probably nothing special. It does not matter if it is special or not. My guess (that may be wrong) that even noveau vs. nvidia may behave differently. I have never gone deeper, since I do not remember any problem with setting DISPLAY=:0 when it was necessary. Driver in use should appear in Xorg.0.log, e.g. (II) modeset(0): [DRI2] DRI driver: i965 I have heard that a display may have several screens (it is not the same as multiple monitors that show different regions of the same display and screen). I have never tried such configuration. Are you referring to multiple desktops? I have that going, for sure. My impression is that multiple screens of a display is not the same as virtual desktops (and not the same as multiple monitors). I am not familiar with X11 protocol so closely. Frankly speaking, I has a hope that somebody will post a proper link. My curiosity is not strong enough yet to filter search engine results myself.
Re: Remove route '169.254.0.0/16 dev ovs-system'
On 26/02/2023 18:18, Geert Stappers wrote: AIUI is systemd-networkd the main player, dhcpcd some helper and NetworkManager for contact with user. First of all, I would not touch dhcpcd.conf for a while. Is it assumed that ovs-system should get its IP address from DHCP? If I understand it correctly, systemd-networkd may send DHCP request itself, it does not need dhcpcd helper. NetworkManager is not only UI, it may manage interfaces. I would not be surprised if more than one tool tries to manage ovs-system. Perhaps some of the following command may shed some light on what really happens: systemctl --failed networkctl list networkctl status ovs-system Last command should show Network File name if the interface is really managed by systemd-networkd. Please, post its content nmcli general status nmcli device nmcli -f CONNECTIONS device show ovs-system ifquery --list and check if /etc/network/interfaces and /etc/network/interfaces.d for ovs-system entries See also https://wiki.debian.org/NetworkConfiguration
unbound and fetchmail (was: Re: Remove route '169.254.0.0/16 dev ovs-system')
On 25/02/2023 19:49, Christoph Brinkhaus wrote: Now there are no messages reported by journald as above. I am curious if fixing unbound and so network-online.target helped to avoid 169.254.x.y address in your case. Can fetchmail work without a kludge you added to achieve some delay? My expectation is that unbound.service may be dropped from Requires= and After= in fetchmail.service, but both fields should have network-online.target. I forgot that debugging of such issues should be started with "systemctl --failed".
Re: unbound and fetchmail (was: Re: Remove route '169.254.0.0/16 dev ovs-system')
On 28/02/2023 17:25, Christoph Brinkhaus wrote: I will just inform about the status. Everything is fine now. A word about systemd-networkd-wait-online: With this service running there has been even a delay of 1-2 seconds when switching from one console to a different one (the consoles when X is not running). I have no idea about that side effect. Does it happen each time or it is getty startup time? In the latter case you may try (for various console numbers) systemd-analyze critical-chain getty@tty1.service Now I have disabled systemd-networkd-wait-online and I have add a comment in the last line of /usr/lib/systemd/system//systemd-networkd.service which is now # 28.2.2023 Also=systemd-networkd-wait-online.service Ideally the root cause of strange behavior should be debugged. Perhaps some dependency should be removed from e.g. multi-user.target or network-online.target. Probably "systemd-analyze critical-chain" may give some hints. Any case I would not touch files in /usr/lib/systemd. It should confuse tools like systemd-delta. It is possible to override specific properties by creating of drop-in snippets in /etc/systemd/system/systemd-networkd.service.d/. See systemd.unit(5), e.g. run systemctl edit systemd-networkd.service [Install] # Clear Also= # Add other values from the original file Also=systemd-networkd.socket Check result systemd-analyze verify systemd-networkd.service I hope, it is a better way to apply your workaround. I can not suggest a better solution for tty delay issue, but I think it exists. "Also=" should affect "systemctl enable systemd-networkd.service" and "disable" commands, so I puzzled why it helps at all.
Re: Forcing dhclient to not ignore tun0 interface when it's available
On 03/03/2023 10:08, Tim Woodall wrote: New to this thread, so might be totally off-piste but openvpn has hooks to run scripts like this: ... This is server side but the route-up/pre-down work client side too. Presumably you can do something here to renew dhcp leases or restore resolv.conf. Perhaps the opposite. dhclient running for enp2s0f0 should detect that VPN is active and to avoid overwriting DNS settings that direct requests to tun0.
Re: Forcing dhclient to not ignore tun0 interface when it's available
On 03/03/2023 13:29, Tim Woodall wrote:
On Fri, 3 Mar 2023, Max Nikulin wrote:
dhclient running for enp2s0f0 should detect that
VPN is active and to avoid overwriting DNS settings that direct
requests to tun0.
The hook can create and delete a file like rhis:
tim@dirac:/etc/dhcp (none)$ cat dhclient-enter-hooks.d/nodnsupdate
make_resolv_conf() {
:
}
I agree that VPN script may add and remove dhclient hook or may write
some file in /run that is read by dhclient hook. They should cooperate
in some way. In more versatile configuration domain resolution may be
per-interface. E.g. hosts from the corporate domain are resolved through
tun0, other sites through enp2s0f0.
Re: unbound and fetchmail (was: Re: Remove route '169.254.0.0/16 dev ovs-system')
On 02/03/2023 22:27, Christoph Brinkhaus wrote: Am Thu, Mar 02, 2023 at 09:26:33PM +0700 schrieb Max Nikulin: On 28/02/2023 17:25, Christoph Brinkhaus wrote: I will just inform about the status. Everything is fine now. A word about systemd-networkd-wait-online: With this service running there has been even a delay of 1-2 seconds when switching from one console to a different one (the consoles when X is not running). I have no idea about that side effect. Does it happen each time or it is getty startup time? In the latter case you may try (for various console numbers) systemd-analyze critical-chaingetty@tty1.service It is happening each time when changing the console. I have no idea which way it may be related to network configuration in general and to 169.254.x.y link local addresses in particular. It is better to start a new thread. - Does journalctl -f show some messages during such delay? - Do you mean that each of [Ctrl+Alt+F3], [Ctrl+Alt+F4], [Ctrl+Alt+F3] hit in sequence cause delay? - Policy Kit may need to adjust permissions to some devices (video, audio, etc.), but 2 seconds is unreasonably long delay.
Re: besides ping/nslookup, any other tools to resolve a hostname?
On 04/03/2023 07:53, Ken Young wrote: Do you know any other way to resolve a hostname by manual on this container? systemd-resolve However I am unsure concerning k8s containers. Some interpreters: python3 -c 'import socket as s, sys; print(s.gethostbyname(sys.argv[1]))' debian.net
Re: what method do you prefer for data transfer between nodes?
On 05/03/2023 15:22, Ken Young wrote: 1. scp Notice that in bookworm (next release) scp uses SFTP protocol. This ssh-based protocol can be used in the current stable release as well. Do not confuse it with ftps that is traditional FTP over TLS.
Re: Looking for "Package Verification with dpkg: Implementation" Document
On 07/03/2023 18:19, Cédric Van Rompay wrote: > This program implements the verification specs defined in the document, "Package Verification with dpkg: Implementation", which is a more complete reference for the verification procedure. ... Any idea which document is this refering to? From search engine results: http://quux.org:70/devel/debian/debsigs.txt Package Verification with dpkg: Implementation John Goerzen Version 5; January 4, 2001
Re: Forcing dhclient to not ignore tun0 interface when it's available
On 06/03/2023 19:17, davenull wrote: On 2023-03-03 06:22, Max Nikulin wrote: Perhaps the opposite. dhclient running for enp2s0f0 should detect that VPN is active and to avoid overwriting DNS settings that direct requests to tun0. Yes, indeed. I want dhclient to NOT overwrite /etc/resolv.conf when VPN is active. OR to use tun05 when it tries to renew the lease ... If anyone has a good documention on how to configure openresolv correctly to use it with openconnect. People suggested openvpn scripts and dhclient hooks in this thread. It should be enough to write a couple of scripts that conditionally update resolv.conf. I am not sure that it is possible to provide configuration that would work out of the box. If you are seeking a ready to use recipe, perhaps you should ask openvpn community. I used network-manager-openconnect-gnome for some time and it was enough to fill some fields in a GUI form for minimal working configuration.
Re: PDF on debian
On 09/03/2023 17:32, Corey Hickman wrote: What's the suggested PDF generator in Debian (without desktop)? And is there a VIM plugin for that? If you use some markup language that can be converted to HTML then there is an alternative to LaTeX workflow: chromium --headless --disable-gpu --print-to-pdf=/tmp/test.pdf file:///tmp/test.html
Re: PDF on debian
On 09/03/2023 22:11, Vincent Lefevre wrote: libreoffice --headless --convert-to pdf file.txt produces a PDF of better quality (fewer spacing issues). I was assuming something like markdown/reStructuredText/asciidoc/etc. instead of plain text. As the last resort :TOhtml vim command to get at least syntax highlighting. Certainly the approach with browser requires a proper CSS for text width, hyphenations, etc.
Re: PDF on debian
On 09/03/2023 21:29, tomas wrote: TeX is perfectly fine as a PDF backend. Especially if you go for the more "modern" variants, like LuaTeX, which grok UTF-8 natively. I am curious if you can provide preamble with font configuration working for most of users for documents including non-latin scripts and emoji. log(😅)=💧log(😄) My experience is that while browsers or office applications implement fallbacks to take missing glyphs from other available fonts out of the box, for LuaLaTeX it is necessary to configure all font fallbacks explicitly. Ideally document should have explicit markup to specify languages (if script is unique then babel can be instructed to detect it and set language implicitly).
Re: Debian 12 - IPv4 blocked without fail2ban & co
On 08/09/2023 04:39, Romain wrote: I can confirm that when this happens, it's the OVH server that fails to send the response to my network. 35 9.862648672 MY_PUBLIC_IP_AT_HOME → 54.38.38.159 ICMP 78 Echo (ping) request id=0x4b30, seq=33150/32385, ttl=1 36 9.862704895 54.38.38.159 → MY_PUBLIC_IP_AT_HOME ICMP 106 Destination unreachable (Port unreachable) Some intermediate firewall may have specific rules for ICMP. It is better to try tcpdump (its capture file may be later loaded to wireshark) for HTTP and SSH packets at 54.38.38.159 and 192.168.0.2 rpi4.home, or if your router allows it, even better at 192.168.0.1 livebox.home.
Re: xclip - how to use without a graphics display for testing on Github workflows
On 10/09/2023 06:36, Tom Browder wrote: We have a Raku module that uses "xclip" during use on a computer with a monitor. We need to test it with Github workflows which does not have a graphics device. Is there any "xclip" option to allow for testing without a graphics devivce without throwing an error? Or any suggestion for a work around? You can create a mock-up and use it instead of real xclip binary.
Re: xclip - how to use without a graphics display for testing on Github workflows
On 10/09/2023 16:44, Tom Browder wrote:
On Sat, Sep 9, 2023 at 21:06 Max Nikulin wrote:
You can create a mock-up and use it instead of real xclip binary.
Sounds interesting, Max, can you show the code?
Unless you need to test subtle issues like
https://github.com/astrand/xclip/issues/20
"Not closing stdout when setting clipboard from stdin"
faced by e.g. tmux users, something simple should be enough:
#!/bin/sh -eu
: "${XCLIP_MOCK_FILE:=$HOME/.xclip-mock}"
while [ $# -gt 0 ]; do
case "$1" in
-i|-in)
;;
-o|--out) exec cat -- "$XCLIP_MOCK_FILE"
;;
-selection|-target) shift
;;
*) # FIXME xclip treats all unknown options at any position as
files
break
;;
esac
shift
done
exec cat -- "$@" >"$XCLIP_MOCK_FILE"
Re: su
On 14/09/2023 17:30, Vincent Lefevre wrote: Yes, XDG_RUNTIME_DIR is problematic, but as "su -" doesn't work in GNU Screen (it yields major display issues), I presume that some environment variables (terminal related?) are still useful. I just have tried it. "su -" preserves TERM=screen.xterm-256color. Behavior is different from pure xterm-256color, but I have no guess what you may consider as major display issues. E.g. vim is even able to determine if background is dark or light and to adjust color scheme accordingly.
Re: using ddrescue on the root partition - boot with / as read-only
On 14/09/2023 19:48, Michael Kjörling wrote: On 14 Sep 2023 12:17 +0200, from vinc...@vinc17.net (Vincent Lefevre): badblocks says that there are 25252 bad blocks. I'm using ddrescue before doing anything else (mainly in case things would go worse), but I would essentially be interested in knowing which files are affected. There's always the brute-force way; for each ordinary file, try to cat it to /dev/null, and log the name of the file if that fails. ddrescue should be a better way to recover data from failing disk. It tries to minimize seeks and may reiterate over fragments failed during earlier passes. It may write session/log file to track data recovered during previous runs. Not only the root partition should be mounted read-only. No partitions from this disk should be mounted at all. Even reads may cause more severe damage. Boot from a live media and run ddrescue from it. Check that no partitions are automatically mounted or used as swap. It is better to postpone badblocks till disk image will be created. If data are really precious then seek for a specialized service.
Re: su
On 14/09/2023 22:26, Vincent Lefevre wrote: I noticed the issue just before the upgrade to bookworm (I wanted to do that for the upgrade). But I can't reproduce it in bookworm. So this may have been an old bug that has been fixed. I do not follow the topic, so I can not attribute changes to particular releases, but certainly I see progress: new features in terminal applications, changes in terminal definitions, adding new ones like tmux flavors. IIRC, this was text positioning issues, perhaps when recalling commands from the history. It may be related to a readline library. E.g. bash does not rely on xterm*VT100.reverseWrap: true BTW, I don't think that the preservation of $TERM is sufficient. Preserving $TERMINFO (when set) is important too. I do not use TERMINFO, however if you need to preserve it than you may use su --whitelist-environment=TERMINFO - My use case is lxc-attach and DISPLAY.
Re: Fresh install, Bookworm, XFCE keeps recreating directories
On 16/09/2023 10:09, Greg Wooledge wrote: Altering the contents of an existing file in ~/.config/ upon login sounds incredibly wrong to me, to the point where I have a hard time believing it's a default behavior. user-dirs.dirs(5) The $HOME/.config/user-dirs.dirs file is a text file that contains the user-specific values for the XDG user dirs. It is created and updated by the xdg-user-dirs-update command. xdg-user-dirs-update(1) On the first run a user-dirs.locale file is created containing the locale that was used for the translation. This is used later by GUI tools like xdg-user-dirs-gtk-update to detect if the locale was changed, letting you to migrate from the old names. So the declared goal is keeping folder names consistent with the current locale.
When a random string is not random (was Re: PATH revisited: one PATH to "rule the [Debian] World")
On 25/09/2023 00:27, Greg Wooledge wrote: On Sun, Sep 24, 2023 at 01:05:32PM -0400, Dan Ritter wrote: Tom Browder wrote: Every time I set up a new host, I have to jump through the hoops trying to get the same PATH for ordinary users as well as root, regardless of how they log in. Reading the man pages doesn't help my old brain with all the caveats. I see the message from Dan Ritter neither on lists.debian.org nor on news.gmane.io. It seems, the person sends messages with fixed Message-ID: <3a6b048e-84db-4a47-969c-b62c826e0...@randomstring.org>. At first I thought it is a mailer bug, but the host name is quite specific. Look at https://lists.debian.org/debian-user/2023/09/threads.html#00175 It is a real mess of unrelated subjects merged into single thread. I suspected that it is Gene who may start a discussion unrelated to original thread, but this case the cause is different. The message from Greg: https://lists.debian.org/debian-user/2023/09/msg00491.html Notice References: Re: usrmerge on root NFS will not be run automatically From: Dan Ritter Due to the In-Reply-To header. Certainly Greg replied to another message, but it was dropped by archivers. So if you see that somebody posted a message to a wrong thread, it is not necessary their fault.
