Re: need google chrome installation instructions that work
On 11/21/18, Patrick Bartek wrote: > On Wed, 21 Nov 2018 11:21:15 -0500 > Greg Wooledge wrote: > >> On Wed, Nov 21, 2018 at 08:02:23AM -0800, Patrick Bartek wrote: >> > I use gdebi (it's in the repos) to install locally saved .deb files. >> > Apt-get won't install local files, that is, not in a repo. >> >> Yes, it will. But you have to supply the filename with a >> leading / or ./ or ../ prefix. >> >> sudo apt-get install ./google-chrome-stable*.deb > > That never worked for me. I've tried. Numerous times. Even with > complete paths, changing to directory where file was, full names, etc., > etc. Both with Wheezy & Stretch. Read somewhere when troubleshooting, > apt-get wouldn't install local .deb. Only way it would was if you > created a local repo and put it in your sources.list. Too much > trouble. That's why I starting using gdebi for those very rare times > I need to. Ditto on that experience for me. Didn't work, but can't remember the (fairly straight up) process to repeat to better understand why not. Had seen the "./" reference on Debian-User in the last week so I tried it while consciously thinking, "WHAT are we doing here?!" In my case, "we" ended up doing nothing because it didn't work. I dropped it and kept moving while thinking I had simply missed some setup detail that would have caused success instead of failure. > Don't know if apt does. Never tried. > > Just occurred to me that since I did customized installs of both Wheezy > and Stretch (No desktop environment, window manager only) starting with > a terminal only system that could be why apt-get won't install > local .deb. Doesn't matter. Gdebi works fine. Have tripped over gdebi a few times in the last year or so. Those sightings were most likely most often as an apt-get suggested download. Will have to give it a try now that this thread has highlighted its usefulness. It's good to have alternative tools when individual Life needs mandate making difficult package usage decisions. We just had that recent thread about playing the escalating dependency game by hand. I broke one my setups k/t that very thing only hours after having commented on that same thread. Fix-missing/fix-broken both shoved a hand in my face while yelling, *Uh-uh, YOU made this mess, YOU clean it up by yourself!* PS I did clean it up, hence the newest debootstrap. :) Cindy :) -- Cindy-Sue Causey Talking Rock, Pickens County, Georgia, USA * runs with duct tape *
Re: VLC doesn't shutdown when closed
On 2018-11-21 8:54 p.m., Cindy-Sue Causey wrote: On 11/21/18, Gary Dale wrote: On 2018-11-21 2:55 a.m., to...@tuxteam.de wrote: On Tue, Nov 20, 2018 at 08:29:18PM -0500, Gary Dale wrote: I'm running Debian/Buster with a Plasma 5 desktop. After I close a VLC session [...] Note that just ending the process doesn't work. I need to send the Kill signal. Not a VLC user here, so take my advice with a fist of salt (or two), but I think you might want to describe more precisely what you mean by "close a VLC session" resp. "end the process". You close the program's main window? Cheers -- tomás Close - click on the close button, which closes the window and should also shut down the application End Process - click on the End Process option in the System Activity monitor after selecting the VLC process. According the manual, this is supposed to send the SIGKILL signal but it doesn't seem to work in this case. Instead I need to send the Kill signal to the process. Cliff's Notes version > jump down to where I said I might have found something. :) The rambling version > More apples and oranges, Thunar file manager was doing that to me a couple years ago to the point it was still alive *after reboots*. That's bad. At some point, I stopped noticing it occurring. Am writing to ask.. is there a toggle switch anywhere, e.g. in Preferences, that gives you the option of leaving it running in the background? I've seen programs do that but can't remember what did. Some browsers have something along that line, for one thing. I just installed Samba and noticed a "background" option for it via its manpages. That wouldn't be something that users necessary knew was occurring because it was about a command line option in Samba's case. One of the players I've used had the option to postpone screensavers until videos finish playing, too. Maybe... well, I don't know. It's nice when the screensaver does NOT kick in while something's playing. But if it's a royal pain for whatever reason, maybe untoggling that might somehow help *if that option exists*. Might have found something. After I wrote the above, I took a quick peek at "apt-cache show vlc" before sending this: "VLC can also be used as a streaming server that duplicates the stream it reads and multicasts them through the network to other clients, or serves them through HTTP." Maybe that has something to do with it not going gentle into that good night? That "show me" there even references an additional plugin package for-r-r-r-r. backgrounding happy Samba. Cindy :) I looked for a keep running option (and VLC does have a lot of options) but couldn't find one.
systemd-networkd and bonding
I'm running Debian/Stretch (AMD64). I'm trying to create a bond between two network devices (currently testing on my laptop but also have a couple of servers I'd like to use it on) following the Debian Wiki at https://wiki.debian.org/Bonding. I got the ifenslave method working as per the first two examples. Then I noticed the systemd-networkd method which looks to be the way of the future. I put the laptop's networking back to it's original settings and removed ifenslave then proceeded with the example, changing the bond device's IP to one that works on my network. The network seems to come up and the bond device has the correct IP but my network doesn't work. Since Mode=802.3ad could have problems, I switched it to active-backup, which worked with ifenslave. Rebooting was slow and when I logged in, I found the network still wasn't working, although ifconfig showed exactly what I thought I should see (the bond device with an IP address and the two slave devices working but without an IP). There's not a lot of online documentation that I've found that doesn't use ifenslave. Has anyone got this to work using systemd-networkd? Also, for my laptop, using dhcp to set the IP makes more sense. The wiki article sets a static IP: [Match] Name=test-lag [Network] Address=192.168.1.13/24 Gateway=192.168.1.1 is how I set it on my laptop. What would that be for a dhcp IP address? Thanks.
Re: VLC doesn't shutdown when closed
On 2018-11-22, Gary Dale wrote: >> > I looked for a keep running option (and VLC does have a lot of options) > but couldn't find one. > If in 'Simple Preferences' under 'Playlists and instances' 'use only one instance when started from file manager' isn't checked the phenomenon has been known to occur (if you're starting VLC from a file manager). If not and on the other hand this appears to be a known bug. How you are closing the app might be pertinent. 'Ctrl Q' after stopping the video? Or some other way? Have you experimented in this regard? -- He used sentences differently from any other prose writer. He always sounded like a slightly drunk man who is very melancholy, who has no illusions about life, who is very strong but whose strength is entirely unnecessary. --Krasznahorkai on Krúdy
Re: need google chrome installation instructions that work
On Wed 21 Nov 2018 at 11:44:02 (-0800), Patrick Bartek wrote: > On Wed, 21 Nov 2018 11:21:15 -0500 > Greg Wooledge wrote: > > On Wed, Nov 21, 2018 at 08:02:23AM -0800, Patrick Bartek wrote: > > > I use gdebi (it's in the repos) to install locally saved .deb files. > > > Apt-get won't install local files, that is, not in a repo. > > > > Yes, it will. But you have to supply the filename with a > > leading / or ./ or ../ prefix. > > > > sudo apt-get install ./google-chrome-stable*.deb > > That never worked for me. I've tried. Numerous times. Even with > complete paths, changing to directory where file was, full names, etc., > etc. Both with Wheezy & Stretch. Read somewhere when troubleshooting, > apt-get wouldn't install local .deb. Only way it would was if you > created a local repo and put it in your sources.list. Too much > trouble. That's why I starting using gdebi for those very rare times > I need to. > > Don't know if apt does. Never tried. It works fine for me, as shown by the output below. However, I notice there's a line: Note, selecting 'xtoolwait' instead of ' … … /xtoolwait_1.3-6.2_amd64.deb' which could suggest that the internals of the .deb file are being taken account of; I haven't studied the output from strace. xtoolwait is a legitimate Debian package, but it hasn't been available since squeeze. I also haven't checked wheezy: all this could postdate wheezy/jessie anyway. For a full path, the output from apt and apt-get is identical (except that apt produces a progress bar, which is later overwritten). - Full path succeeds: ~# apt-get install /home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb Reading package lists... Done Building dependency tree Reading state information... Done Note, selecting 'xtoolwait' instead of '/home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb' The following NEW packages will be installed: xtoolwait 0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded. Need to get 0 B/13.1 kB of archives. After this operation, 77.8 kB of additional disk space will be used. Get:1 /home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb xtoolwait amd64 1.3-6.2 [13.1 kB] Retrieving bug reports... Done Parsing Found/Fixed information... Done Selecting previously unselected package xtoolwait. (Reading database ... 261615 files and directories currently installed.) Preparing to unpack .../xtoolwait_1.3-6.2_amd64.deb ... Unpacking xtoolwait (1.3-6.2) ... Setting up xtoolwait (1.3-6.2) ... Processing triggers for man-db (2.7.6.1-2) ... Scanning processes... Scanning linux images... Running kernel seems to be up-to-date. No services need to be restarted. No containers need to be restarted. No user sessions are running outdated binaries. N: Download is performed unsandboxed as root as file '/home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb' couldn't be accessed by user '_apt'. - pkgAcquire::Run (13: Permission denied) ~# - Filename alone fails: /home/debian/squeeze# apt-get install xtoolwait_1.3-6.2_amd64.deb Reading package lists... Done Building dependency tree Reading state information... Done E: Unable to locate package xtoolwait_1.3-6.2_amd64.deb E: Couldn't find any package by glob 'xtoolwait_1.3-6.2_amd64.deb' E: Couldn't find any package by regex 'xtoolwait_1.3-6.2_amd64.deb' 100 /home/debian/squeeze# - ./ path succeeds (and ../ likewise): /home/debian/squeeze# apt-get install ./xtoolwait_1.3-6.2_amd64.deb Reading package lists... Done Building dependency tree Reading state information... Done Note, selecting 'xtoolwait' instead of './xtoolwait_1.3-6.2_amd64.deb' The following NEW packages will be installed: xtoolwait 0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded. Need to get 0 B/13.1 kB of archives. After this operation, 77.8 kB of additional disk space will be used. Get:1 /home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb xtoolwait amd64 1.3-6.2 [13.1 kB] Retrieving bug reports... Done Parsing Found/Fixed information... Done Selecting previously unselected package xtoolwait. (Reading database ... 261615 files and directories currently installed.) Preparing to unpack .../xtoolwait_1.3-6.2_amd64.deb ... Unpacking xtoolwait (1.3-6.2) ... Setting up xtoolwait (1.3-6.2) ... Processing triggers for man-db (2.7.6.1-2) ... Scanning processes... Scanning linux images... Running kernel seems to be up-to-date. No services need to be restarted. No containers need to be restarted. No user sessions are running outdated binaries. N: Download is performed unsandboxed as root as file '/home/debian/squeeze/xtoolwait_1.3-6.2_amd64.deb' couldn't be accessed by user '_apt'. - pkgAcquire::Run (13: Permission denied)
Re: systemd-networkd and bonding
On 11/22/2018 4:11 PM, Gary Dale wrote: > I'm running Debian/Stretch (AMD64). > > I'm trying to create a bond between two network devices (currently > testing on my laptop but also have a couple of servers I'd like to use > it on) following the Debian Wiki at https://wiki.debian.org/Bonding. > > I got the ifenslave method working as per the first two examples. Then I > noticed the systemd-networkd method which looks to be the way of the > future. I put the laptop's networking back to it's original settings and > removed ifenslave then proceeded with the example, changing the bond > device's IP to one that works on my network. > > The network seems to come up and the bond device has the correct IP but > my network doesn't work. Since Mode=802.3ad could have problems, I > switched it to active-backup, which worked with ifenslave. Rebooting was > slow and when I logged in, I found the network still wasn't working, > although ifconfig showed exactly what I thought I should see (the bond > device with an IP address and the two slave devices working but without > an IP). > > There's not a lot of online documentation that I've found that doesn't > use ifenslave. Has anyone got this to work using systemd-networkd? > Not tested: https://www.reversengineered.com/2014/08/21/setting-up-bonding-in-systemd/ > Also, for my laptop, using dhcp to set the IP makes more sense. The wiki > article sets a static IP: > > [Match] > Name=test-lag > [Network] > Address=192.168.1.13/24 > Gateway=192.168.1.1 > [Network] DHCP=yes See "Example 2. DHCP on ethernet links" and the "Network" section "DHCP=" at the following URL : https://www.freedesktop.org/software/systemd/man/systemd.network.html -- John Doe
Re: compiz: Migrating to compiz-reloaded?
Hello, Samuel Thibault, le jeu. 30 août 2018 17:26:27 +0200, a ecrit: > The launchpad upstream for compiz has switched to light maintenance mode > and will not make further development since Ubuntu has stopped using it. > > How do people feel about switching to compiz-reloaded? > > https://gitlab.com/compiz No reaction, so we moved on :) Compiz-reloaded 0.8.14 is currently in sid, and 0.8.16 is currently in experimental, up for your testing :) For debian-accessibility: a very interesting feature of 0.8.16 is the ability of ezoom to track focus events, which makes it a very effective screen magnifier. Samuel
Unable to boot BIOS/GPT system.
Hello, I am trying to install Debian on a 4TB external HDD. Since I want to boot the HDD (also) on old systems which support only BIOS, and since the HDD is larger than 2TB, I decided to go with GPT partition table and BIOS boot (or, in other words: BIOS/GTP). I partitioned the HDD accordingly (see the details below), and installed Debian. During a pre-installation configuration the installer recognized BIOS boot partition, I configured the root partition as ext4 and installed grub into the disk's MBR. But when (after an installation) I am trying to boot the HDD, I receive an error message: 'error: unknown filesystem' and drop into a grub rescue mode. What is the problem here and how it can be solved? It seems as though grub's Stage 1.5 loader (that is - core.img, which is supposed to be embedded in BIOS boot partition) doesn't contain modules needed to read ext4 filesystem, which seems very strange. Following are a bit more details about the partitions: I used the following command to partition the HDD: parted -s /dev/sdb mklabel gpt mkpart BIOS_boot ext2 1MiB 2MiB \ mkpart swap linux-swap 2MiB 16000MiB mkpart data ext2 16000MiB 100% \ set 1 bios_grub [root@arch]# parted /dev/sdb unit MiB print Model: WD Elements 25A1 (scsi) Disk /dev/sdb: 3815415MiB Sector size (logical/physical): 512B/4096B Partition Table: gpt Disk Flags: Number Start End SizeFile system Name Flags 1 1.00MiB 2.00MiB 1.00MiB bios_grub 2 2.00MiB 16000MiB15998MiBlinux-swap(v1) 3 16000MiB 3815414MiB 3799414MiB Debian's installation was done on x64 BIOS-only system.
Install openssh-server jessie version deb package on stretch
Hi, I have compatibility issues with the latest version of openssh-server and an old dropbear client, the dopbear client stops at preauth *ov 22 14:34:03 * *myhostname **sshd[3905]: debug1: Client protocol version 2.0; client software version dropbear_0.46* *Nov 22 14:34:03 **myhostname** sshd[3905]: debug1: no match: dropbear_0.46* Nov 22 14:34:03 myhostname sshd[3905]: debug1: Local version string SSH-2.0-OpenSSH_7.4p1 Debian-10+deb9u4 Nov 22 14:34:03 myhostname sshd[3905]: debug1: Enabling compatibility mode for protocol 2.0 Nov 22 14:34:03 myhostname sshd[3905]: debug2: fd 3 setting O_NONBLOCK Nov 22 14:34:03 myhostname sshd[3905]: debug2: Network child is on pid 3906 Nov 22 14:34:03 myhostname sshd[3905]: debug3: preauth child monitor started Nov 22 14:34:03 myhostname sshd[3905]: debug3: privsep user:group 106:65534 [preauth] Nov 22 14:34:03 myhostname sshd[3905]: debug1: permanently_set_uid: 106/65534 [preauth] Nov 22 14:34:03 myhostname sshd[3905]: debug1: list_hostkey_types: ssh-rsa,rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth] Nov 22 14:34:03 myhostname sshd[3905]: debug3: send packet: type 20 [preauth] *Nov 22 14:34:03 myhostname sshd[3905]: debug1: SSH2_MSG_KEXINIT sent [preauth]* I'm thinking about installing the previous version of the package (Jessie) http://ftp.it.debian.org/debian/pool/main/o/openssh/openssh-server_7.9p1-4_amd64.deb Do you say that it is possible? Thanks
Re: Install openssh-server jessie version deb package on stretch
On Thu, Nov 22, 2018 at 07:32:07PM +0100, owl...@gmail.com wrote: >Hi, I have compatibility issues with the latest version of openssh-server >and an old dropbear client, the dopbear client stops at preauth >ov 22 14:34:03 myhostname sshd[3905]: debug1: Client protocol version >2.0; client software version dropbear_0.46 >Nov 22 14:34:03 myhostname sshd[3905]: debug1: no match: dropbear_0.46 >Nov 22 14:34:03 myhostname sshd[3905]: debug1: Local version string >SSH-2.0-OpenSSH_7.4p1 Debian-10+deb9u4 >Nov 22 14:34:03 myhostname sshd[3905]: debug1: Enabling compatibility mode >for protocol 2.0 >Nov 22 14:34:03 myhostname sshd[3905]: debug2: fd 3 setting O_NONBLOCK >Nov 22 14:34:03 myhostname sshd[3905]: debug2: Network child is on pid >3906 >Nov 22 14:34:03 myhostname sshd[3905]: debug3: preauth child monitor >started >Nov 22 14:34:03 myhostname sshd[3905]: debug3: privsep user:group >106:65534 [preauth] >Nov 22 14:34:03 myhostname sshd[3905]: debug1: permanently_set_uid: >106/65534 [preauth] >Nov 22 14:34:03 myhostname sshd[3905]: debug1: list_hostkey_types: >ssh-rsa,rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth] >Nov 22 14:34:03 myhostname sshd[3905]: debug3: send packet: type 20 >[preauth] >Nov 22 14:34:03 myhostname sshd[3905]: debug1: SSH2_MSG_KEXINIT sent >[preauth] >I'm thinking about installing the previous version of the package (Jessie) > > [1]http://ftp.it.debian.org/debian/pool/main/o/openssh/openssh-server_7.9p1-4_amd64.deb >Do you say that it is possible? >Thanks > That is actually a terrible idea. You are better off editing /etc/sshd_config and enabling appropriate key exchange and cipher algorithms that are compatible with the old dropbear client. Given the potential security issues there, a beter approach is to instead create a copy of the current configuration, make the necessary changes to be compatible with dropbear, then run two sshd instances. Make the one with the weak algorithms only accessible to the IP from which the dropbear connection will initiate (you can do this in your system firewall) and then make it only accessible to the specific user (you can do this with an AllowUsers directive in that instances sshd_config). The other instance can remain accessible as you currently have it with no degradation of security. You will also need to decide which instance will run on which ports, since both cannot occupy the same port. Alternately, if the machine has multiple IP addresses, the two instances can be on the same port bound to different addresses. Regards, -Roberto -- Roberto C. Sánchez
Re: Unable to boot BIOS/GPT system.
On 11/22/18 8:24 PM, Daniel Fishman wrote: > Hello, > > I am trying to install Debian on a 4TB external HDD. Since I want > to boot the HDD (also) on old systems which support only BIOS, > and since the HDD is larger than 2TB, I decided to go with GPT > partition table and BIOS boot (or, in other words: BIOS/GTP). > I partitioned the HDD accordingly (see the details below), and > installed Debian. During a pre-installation configuration the > installer recognized BIOS boot partition, I configured the root > partition as ext4 and installed grub into the disk's MBR. But > when (after an installation) I am trying to boot the HDD, I receive > an error message: 'error: unknown filesystem' and drop into a > grub rescue mode. > > What is the problem here and how it can be solved? It seems as > though grub's Stage 1.5 loader (that is - core.img, which is > supposed to be embedded in BIOS boot partition) doesn't contain > modules needed to read ext4 filesystem, which seems very strange. > > > Following are a bit more details about the partitions: > > I used the following command to partition the HDD: > > parted -s /dev/sdb mklabel gpt mkpart BIOS_boot ext2 1MiB 2MiB \ > mkpart swap linux-swap 2MiB 16000MiB mkpart data ext2 16000MiB 100% \ > set 1 bios_grub > > [root@arch]# parted /dev/sdb unit MiB print > Model: WD Elements 25A1 (scsi) > Disk /dev/sdb: 3815415MiB > Sector size (logical/physical): 512B/4096B > Partition Table: gpt > Disk Flags: > > Number Start End Size File system Name Flags > 1 1.00MiB 2.00MiB 1.00MiB bios_grub > 2 2.00MiB 16000MiB 15998MiB linux-swap(v1) > 3 16000MiB 3815414MiB 3799414MiB > > > Debian's installation was done on x64 BIOS-only system. > Hi Daniel, in case of BIOS and GPT you have to create a partition for second stage of GRUB. If I can remember correctly its size should be less than 50MB. HTH Kind regards Georgi
Re: Unable to boot BIOS/GPT system.
Hi Daniel, in case of BIOS and GPT you have to create a partition for second stage of GRUB. If I can remember correctly its size should be less than 50MB. HTH Kind regards Georgi I created it - this is the first partition (the one with bios_grub flag). I did things similarly to the description here: https://bit.ly/2utVdNr
Re: Unable to boot BIOS/GPT system.
quant...@gmail.com wrote: > >I am trying to install Debian on a 4TB external HDD. Since I want >to boot the HDD (also) on old systems which support only BIOS, >and since the HDD is larger than 2TB, I decided to go with GPT >partition table and BIOS boot (or, in other words: BIOS/GTP). >I partitioned the HDD accordingly (see the details below), and >installed Debian. During a pre-installation configuration the >installer recognized BIOS boot partition, I configured the root >partition as ext4 and installed grub into the disk's MBR. But >when (after an installation) I am trying to boot the HDD, I receive >an error message: 'error: unknown filesystem' and drop into a >grub rescue mode. > >What is the problem here and how it can be solved? It seems as >though grub's Stage 1.5 loader (that is - core.img, which is >supposed to be embedded in BIOS boot partition) doesn't contain >modules needed to read ext4 filesystem, which seems very strange. ... >Number Start End SizeFile system Name Flags > 1 1.00MiB 2.00MiB 1.00MiB bios_grub > 2 2.00MiB 16000MiB15998MiBlinux-swap(v1) > 3 16000MiB 3815414MiB 3799414MiB I'd expect that Grub won't be able to reliably work with that setup - partition 3 goes past the 2TB mark, so the BIOS won't be able to map it properly. Don't forget, when grub is reading the disk all it can rely on are BIOS calls. Add yourself a small-ish /boot partition first and you may be OK. -- Steve McIntyre, Cambridge, UK.st...@einval.com "Further comment on how I feel about IBM will appear once I've worked out whether they're being malicious or incompetent. Capital letters are forecast." Matthew Garrett, http://www.livejournal.com/users/mjg59/30675.html
Re: Unable to boot BIOS/GPT system.
On 11/22/18 9:08 PM, Daniel Fishman wrote: > >> >> Hi Daniel, >> >> in case of BIOS and GPT you have to create a partition for second stage >> of GRUB. If I can remember correctly its size should be less than 50MB. >> >> HTH >> >> Kind regards >> Georgi > > I created it - this is the first partition (the one with bios_grub flag). > I did things similarly to the description here: https://bit.ly/2utVdNr Can you try to create all the partitions through Debian's installer. I have a suspicion that partition type may be wrong for BIOS partition. You have specified ext2 as file system for BIOS grub which may mean Linux partition instead of BIOS grub partition (ef02 BIOS boot).
Re: Install openssh-server jessie version deb package on stretch
Thanks Roberto, I have tried also the latest dropbear server but this is incompatible too Do you have idea how can I find appropriate key exchange and cipher algorithms? Il giorno gio 22 nov 2018 alle ore 19:42 Roberto C. Sánchez ha scritto: > > On Thu, Nov 22, 2018 at 07:32:07PM +0100, owl...@gmail.com wrote: > >Hi, I have compatibility issues with the latest version of openssh-server > >and an old dropbear client, the dopbear client stops at preauth > >ov 22 14:34:03 myhostname sshd[3905]: debug1: Client protocol version > >2.0; client software version dropbear_0.46 > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: no match: dropbear_0.46 > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: Local version string > >SSH-2.0-OpenSSH_7.4p1 Debian-10+deb9u4 > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: Enabling compatibility > > mode > >for protocol 2.0 > >Nov 22 14:34:03 myhostname sshd[3905]: debug2: fd 3 setting O_NONBLOCK > >Nov 22 14:34:03 myhostname sshd[3905]: debug2: Network child is on pid > >3906 > >Nov 22 14:34:03 myhostname sshd[3905]: debug3: preauth child monitor > >started > >Nov 22 14:34:03 myhostname sshd[3905]: debug3: privsep user:group > >106:65534 [preauth] > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: permanently_set_uid: > >106/65534 [preauth] > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: list_hostkey_types: > >ssh-rsa,rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth] > >Nov 22 14:34:03 myhostname sshd[3905]: debug3: send packet: type 20 > >[preauth] > >Nov 22 14:34:03 myhostname sshd[3905]: debug1: SSH2_MSG_KEXINIT sent > >[preauth] > >I'm thinking about installing the previous version of the package > > (Jessie) > > > > [1]http://ftp.it.debian.org/debian/pool/main/o/openssh/openssh-server_7.9p1-4_amd64.deb > >Do you say that it is possible? > >Thanks > > > > That is actually a terrible idea. > > You are better off editing /etc/sshd_config and enabling appropriate key > exchange and cipher algorithms that are compatible with the old dropbear > client. > > Given the potential security issues there, a beter approach is to > instead create a copy of the current configuration, make the necessary > changes to be compatible with dropbear, then run two sshd instances. > Make the one with the weak algorithms only accessible to the IP from > which the dropbear connection will initiate (you can do this in your > system firewall) and then make it only accessible to the specific user > (you can do this with an AllowUsers directive in that instances > sshd_config). The other instance can remain accessible as you currently > have it with no degradation of security. > > You will also need to decide which instance will run on which ports, > since both cannot occupy the same port. Alternately, if the machine has > multiple IP addresses, the two instances can be on the same port bound > to different addresses. > > Regards, > > -Roberto > > -- > Roberto C. Sánchez >
compiz: Migrating to compiz-reloaded?
Hello, On 22/11/2018 19:08, Samuel Thibault wrote: > For debian-accessibility: a very interesting feature of 0.8.16 is the > ability of ezoom to track focus events, which makes it a very effective > screen magnifier. I did set that in Slint with Mate 1.18, now the window moves when I type or use the arrow keys so that the cursor stays on the middle of the screen (as Alex showed me at a Samedi du Libre) if I zoom. Very nice, thanks Samuel! I attach my ~/config/compiz/compizconfig/Desktop.ini. I have done some settings including to allow compiz to manage the windows, but it's still WIP, so proposals for enhancements are welcome. Comments: 1. I had to modify the key bindings for Initiate the Magnifier: Alt+m => Alt+o else it conflicted with Toggle Window Negative 2. I modified the keybindings of the Switcher to be able to switch between the panels and the desktop Best, Didier [core] as_active_plugins = core;decoration;colorfilter;resize;neg;glib;mousepoll;matecompat;obs;move;focuspoll;showmouse;ezoom;mag;switcher; [zoom] s0_follow_focus = true [ezoom] s0_follow_focus = true [switcher] as_prev_key = Disabled as_next_panel_key = Tab [mag] as_initiate = o
Re: Install openssh-server jessie version deb package on stretch
On Thu, Nov 22, 2018 at 08:40:52PM +0100, owl...@gmail.com wrote: > Thanks Roberto, > > I have tried also the latest dropbear server but this is incompatible too > > Do you have idea how can I find appropriate key exchange and cipher > algorithms? > Please don't top post. It is considered impolite. I do not use dropbear, but I would expect its documentation should provide a list of supported key exchange and cipher algorithms. You could then look at the output of 'ssh -Q kex' and 'ssh -Q cipher' for lists of key exchange and cipher algorithms supported by ssh on your system, respectively. I also forgot to mention in my earlier reply that you might need to specify weaker MAC algorithms. Available MAC algorithms can be obtained with 'ssh -Q mac'. Regards, -Roberto -- Roberto C. Sánchez
Re: Install openssh-server jessie version deb package on stretch
Il giorno gio 22 nov 2018 alle ore 20:48 Roberto C. Sánchez ha scritto: > > On Thu, Nov 22, 2018 at 08:40:52PM +0100, owl...@gmail.com wrote: > > Thanks Roberto, > > > > I have tried also the latest dropbear server but this is incompatible too > > > > Do you have idea how can I find appropriate key exchange and cipher > > algorithms? > > > Please don't top post. It is considered impolite. > > I do not use dropbear, but I would expect its documentation should > provide a list of supported key exchange and cipher algorithms. You > could then look at the output of 'ssh -Q kex' and 'ssh -Q cipher' for > lists of key exchange and cipher algorithms supported by ssh on your > system, respectively. > > I also forgot to mention in my earlier reply that you might need to > specify weaker MAC algorithms. Available MAC algorithms can be obtained > with 'ssh -Q mac'. > > Regards, > > -Roberto > > -- > Roberto C. Sánchez > Thanks Roberto, tomorrow ill'try
Re: Unable to boot BIOS/GPT system.
I'd expect that Grub won't be able to reliably work with that setup - partition 3 goes past the 2TB mark, so the BIOS won't be able to map it properly. Don't forget, when grub is reading the disk all it can rely on are BIOS calls. Add yourself a small-ish /boot partition first and you may be OK. Thanks, indeed this was the problem - after adding a boot partition the HDD now boots on BIOS-only machines. Unfortunately, it does not boot on UEFI machines. Seems that I have to start experimenting with hybrid UEFI+BIOS boot...
Re: Unable to boot BIOS/GPT system.
Can you try to create all the partitions through Debian's installer. I have a suspicion that partition type may be wrong for BIOS partition. You have specified ext2 as file system for BIOS grub which may mean Linux partition instead of BIOS grub partition (ef02 BIOS boot). In the end, it was actually the lack of /boot partition that caused the problem (as Steve McIntyre explained in another message in this thread).
Re: VLC doesn't shutdown when closed
On 2018-11-22 10:26 a.m., Curt wrote: On 2018-11-22, Gary Dale wrote: I looked for a keep running option (and VLC does have a lot of options) but couldn't find one. If in 'Simple Preferences' under 'Playlists and instances' 'use only one instance when started from file manager' isn't checked the phenomenon has been known to occur (if you're starting VLC from a file manager). I normally start it from Dolphin and don't have that option checked because I sometimes need to compare two videos. If not and on the other hand this appears to be a known bug. How you are closing the app might be pertinent. 'Ctrl Q' after stopping the video? Or some other way? Have you experimented in this regard? I normally just click the close "X" on the top-right. I'm lazy so the video probably isn't stopped. I haven't experimented with other methods because I don't consider a media player staying open in the background to be anything but a bug. I have tried killall vlc but I note that fails to work, confirming my suspicion that this is a bug. A properly behaving program should respond to normal terminate requests.
Re: VLC doesn't shutdown when closed
On 2018-11-22 10:26 a.m., Curt wrote: On 2018-11-22, Gary Dale wrote: I looked for a keep running option (and VLC does have a lot of options) but couldn't find one. If in 'Simple Preferences' under 'Playlists and instances' 'use only one instance when started from file manager' isn't checked the phenomenon has been known to occur (if you're starting VLC from a file manager). If not and on the other hand this appears to be a known bug. How you are closing the app might be pertinent. 'Ctrl Q' after stopping the video? Or some other way? Have you experimented in this regard? Interesting - if I press the stop button on the player, vlc terminates properly.
Re: Unable to boot BIOS/GPT system.
Le 22/11/2018 à 20:19, Steve McIntyre a écrit : quant...@gmail.com wrote: I am trying to install Debian on a 4TB external HDD. Since I want to boot the HDD (also) on old systems which support only BIOS, and since the HDD is larger than 2TB, I decided to go with GPT partition table and BIOS boot (or, in other words: BIOS/GTP). I partitioned the HDD accordingly (see the details below), and installed Debian. During a pre-installation configuration the installer recognized BIOS boot partition, I configured the root partition as ext4 and installed grub into the disk's MBR. But when (after an installation) I am trying to boot the HDD, I receive an error message: 'error: unknown filesystem' and drop into a grub rescue mode. What is the problem here and how it can be solved? It seems as though grub's Stage 1.5 loader (that is - core.img, which is supposed to be embedded in BIOS boot partition) doesn't contain modules needed to read ext4 filesystem, which seems very strange. ... Number Start End SizeFile system Name Flags 1 1.00MiB 2.00MiB 1.00MiB bios_grub 2 2.00MiB 16000MiB15998MiBlinux-swap(v1) 3 16000MiB 3815414MiB 3799414MiB I'd expect that Grub won't be able to reliably work with that setup - partition 3 goes past the 2TB mark, so the BIOS won't be able to map it properly. Don't forget, when grub is reading the disk all it can rely on are BIOS calls. Add yourself a small-ish /boot partition first and you may be OK. The limit of EDD BIOS calls which uses 64 bit LBA addressing is way beyond 2 TiB. But that specific BIOS implementation may be broken and unable to read beyond that limit. So indeed a boot or root partition at the beginning of the disk may help.
grub tries to read outside of a disk.
Hello, I am trying to install Debian on a 4TB external HDD, and want to be able to boot it both on UEFI and BIOS systems. Using a simple BIOS/GPT configuration (GPT partitioning with BIOS boot) didn't work out - after receiving some important tips from this mailing list about problems that I had with this configuration I succeeded to boot on BIOS system, but not on UEFI system. Therefore I tried to combine BIOS+UEFI boots by creating both EFI and BIOS boot partitions, and then installing grub into both. Here is how the HDD was partitioned: [root@arch]# parted /dev/sdb print Model: WD Elements 25A1 (scsi) Disk /dev/sdb: 4001GB Sector size (logical/physical): 512B/4096B Partition Table: gpt Disk Flags: Number Start End SizeFile system Name Flags 1 1049kB 2097kB 1049kBbios_grub 2 2097kB 202MB 200MB fat32 boot, esp 3 202MB 714MB 512MB ext4 4 714MB 8714MB 8000MB linux-swap(v1) 5 8714MB 4001GB 3992GB ext4 Partition 1 is intended for BIOS boot partition (for Stage 1.5 loader), partition 2 is the EFI partition, partition 3 is intended for the /boot (as was suggested previously on this mailing list) so that grub-related files will be located inside the 2TB limit and therefore be accessible to Stage 1.5 loader in BIOS mode, partition 4 is a swap, and 5 is the /. I performed Debian installation on a UEFI machine, and after an installation succeeded to boot the HDD. After that I installed grub into MBR: grub-install --target=i386-pc --recheck /dev/sdb Then I tried to boot on BIOS machine, and got the following error from grub: "error: attempt to read or write outside of disk 'hd0'" and was dropped into grub's rescue shell. I tried the following command: ls (hd0,gpt3)/ which listed the expected contents of the /boot partition. Then: set root=(hd0,gpt3) set prefix=(hd0,gpt3)/grub insmod normal and again I get the error message "error: attempt to read or write outside of disk 'hd0'" As far as I could find, usually this kind of an error happens when grub's files are located on a partition which is far from the beginning of a disk, but in my case this is not so. And since grub can list contents of a partition from a rescue shell, then it has access to the file system. Then what is the problem here?
Re: systemd-networkd and bonding
On 2018-11-22 11:02 a.m., john doe wrote: On 11/22/2018 4:11 PM, Gary Dale wrote: I'm running Debian/Stretch (AMD64). I'm trying to create a bond between two network devices (currently testing on my laptop but also have a couple of servers I'd like to use it on) following the Debian Wiki at https://wiki.debian.org/Bonding. I got the ifenslave method working as per the first two examples. Then I noticed the systemd-networkd method which looks to be the way of the future. I put the laptop's networking back to it's original settings and removed ifenslave then proceeded with the example, changing the bond device's IP to one that works on my network. The network seems to come up and the bond device has the correct IP but my network doesn't work. Since Mode=802.3ad could have problems, I switched it to active-backup, which worked with ifenslave. Rebooting was slow and when I logged in, I found the network still wasn't working, although ifconfig showed exactly what I thought I should see (the bond device with an IP address and the two slave devices working but without an IP). There's not a lot of online documentation that I've found that doesn't use ifenslave. Has anyone got this to work using systemd-networkd? Not tested: https://www.reversengineered.com/2014/08/21/setting-up-bonding-in-systemd/ Also, for my laptop, using dhcp to set the IP makes more sense. The wiki article sets a static IP: [Match] Name=test-lag [Network] Address=192.168.1.13/24 Gateway=192.168.1.1 [Network] DHCP=yes See "Example 2. DHCP on ethernet links" and the "Network" section "DHCP=" at the following URL : https://www.freedesktop.org/software/systemd/man/systemd.network.html Thanks. It took a different approach and used different files than the Debian Wiki article but it did the trick. My laptop setup currently uses some ideas from both (I have .network files using the pci addresses for the laptop network hardware that I refer to in the .netdev file). It's now running with dhcp and Mode=mode802.3ad. One thing I did notice is that I did need a reboot in at least one point to get things working properly. Simply restarting the systemd-networkd service didn't do the trick.
Re: systemd-networkd and bonding
On 11/23/2018 5:23 AM, Gary Dale wrote: > On 2018-11-22 11:02 a.m., john doe wrote: >> On 11/22/2018 4:11 PM, Gary Dale wrote: >>> I'm running Debian/Stretch (AMD64). >>> >>> I'm trying to create a bond between two network devices (currently >>> testing on my laptop but also have a couple of servers I'd like to use >>> it on) following the Debian Wiki at https://wiki.debian.org/Bonding. >>> >>> I got the ifenslave method working as per the first two examples. Then I >>> noticed the systemd-networkd method which looks to be the way of the >>> future. I put the laptop's networking back to it's original settings and >>> removed ifenslave then proceeded with the example, changing the bond >>> device's IP to one that works on my network. >>> >>> The network seems to come up and the bond device has the correct IP but >>> my network doesn't work. Since Mode=802.3ad could have problems, I >>> switched it to active-backup, which worked with ifenslave. Rebooting was >>> slow and when I logged in, I found the network still wasn't working, >>> although ifconfig showed exactly what I thought I should see (the bond >>> device with an IP address and the two slave devices working but without >>> an IP). >>> >>> There's not a lot of online documentation that I've found that doesn't >>> use ifenslave. Has anyone got this to work using systemd-networkd? >>> >> Not tested: >> >> https://www.reversengineered.com/2014/08/21/setting-up-bonding-in-systemd/ >> >> >>> Also, for my laptop, using dhcp to set the IP makes more sense. The wiki >>> article sets a static IP: >>> >>> [Match] >>> Name=test-lag >>> [Network] >>> Address=192.168.1.13/24 >>> Gateway=192.168.1.1 >>> >> [Network] >> DHCP=yes >> >> See "Example 2. DHCP on ethernet links" and the "Network" section >> "DHCP=" at the following URL : >> >> https://www.freedesktop.org/software/systemd/man/systemd.network.html >> > Thanks. It took a different approach and used different files than the > Debian Wiki article but it did the trick. My laptop setup currently uses > some ideas from both (I have .network files using the pci addresses for > the laptop network hardware that I refer to in the .netdev file). It's > now running with dhcp and Mode=mode802.3ad. > Glad to hear that you got it working. > One thing I did notice is that I did need a reboot in at least one point > to get things working properly. Simply restarting the systemd-networkd > service didn't do the trick. > This is expected: From: https://www.freedesktop.org/software/systemd/man/systemd-networkd.service.html "This also means that when configuration is updated and systemd-networkd is restarted, netdev interfaces for which configuration was removed will not be dropped, and may need to be cleaned up manually." -- John Doe
Re: Unable to boot BIOS/GPT system.
Le 22/11/2018 à 21:24, Daniel Fishman a écrit :
I'd expect that Grub won't be able to reliably work with that setup -
partition 3 goes past the 2TB mark, so the BIOS won't be able to map
it properly. Don't forget, when grub is reading the disk all it can
rely on are BIOS calls. Add yourself a small-ish /boot partition first
and you may be OK.
Thanks, indeed this was the problem - after adding a boot partition
the HDD now boots on BIOS-only machines.
So these machines have a broken BIOS implementation.
Unfortunately, it does not boot on UEFI machines.
Of course not. EFI boot requires a FAT-formatted EFI partition and
grub-efi-{amd64|ia32} depending on the firmware architecture (not the CPU).
If the EFI firmware can boot in legacy BIOS compatibility mode, it may
require to set the boot flag on the protective GPT partition entry in
the protective MBR. Use parted disk_set pmbr_boot or fdisk -t dos.
