Re: Wayland vs X
Dear Kevin, Kevin Exton writes: > I tried Wayland some years ago now (might have been when they first > trialled it in Ubuntu) but decided not to stick with it. Well i don't know my login desktop what it is. Always i use default values. Currently i'm using Debian 11 Bullseye under Chromebook. That seems not X11, maybe... And i am happy with the default values. #+begin_src text soyeomul@penguin:~$ lsb_release -d Description:Debian GNU/Linux 11 (bullseye) soyeomul@penguin:~$ env | grep wayland WAYLAND_DISPLAY_LOW_DENSITY=wayland-1 WAYLAND_DISPLAY=wayland-0 #+end_src Sincerely, Linux fan Byung-Hee -- ^고맙습니다 _和合團結_ 감사합니다_^))//
https authorisation server?
I'm working on providing an imap proxy on my LAN gateway and it seems nginx module ngx_mail will do the job nicely. The problem is the module mandates an http authorisation server. Ideally I would run the authorisation server on my internal mail machine and it would use PAM or suchlike to look up mail usernames and passwords rather than me having to maintain separate user/password files. What candidates are there for a standalone https authorisation server? Or a plug in module for an nginx instance running on the internal mail server? Thanks, -- Jeremy OpenPGP_signature Description: OpenPGP digital signature
Re: packages built with golang
On 3/15/22, Dan Ritter wrote: > Cousin Stanley wrote: >> Cousin Stanley wrote : >> > The data is already on your system, so >> > there's no transmission happening. >> >> I do not understand this. > ... > >> Does the Debian package manager >> really download package information >> for all ~59,000 avaiilabel packages >> in anticipation that users will need it >> at sometime in the future ? >> >> That would be suprising to me. > > Now you are surprised, and informed. My linux-image-amd64 upgrade got mangled the other day so I was in "man apt-get" and apt-cache looking for ways to coerce success (beyond apt's recommended "dpkg --configure -a"). Found this and had intended to eventually post here anyway: apt-cache stats Thought it might prove of interest for others, too, with respect to seeing real numbers about the amazing volume of packages all interacting together under Debian's hood. Concerning this thread, the multi-megabyte sizes of the files under /var/lib/apt/lists reflect the package numbers found via "apt-cache stats". I can still remember my first ah-ha moment upon opening up one of those files and peeking in out of curiosity years ago.. Cindy :) -- Talking Rock, Pickens County, Georgia, USA * runs with birdseed *
Re: Archiving content of a directory on a DVD-R.
From: "Thomas Schmitt" Date: Tue, 15 Mar 2022 22:27:37 +0100 > How old is it ? Nameplate is marked "September 2003". > I find this model mentioned as early as 2003. DVD burning was a novelty > back then. Consistent. > You may ask it by > > xorriso -outdev /dev/sr0 -list_profiles out Drive current: -outdev '/dev/sr0' Media current: CD-R Media status : is blank Media summary: 0 sessions, 0 data blocks, 0 data, 703m free Drive current: -outdev '/dev/sr0' Drive type : vendor 'PLEXTOR' product 'DVDR PX-708A' revision '1.06' Profile : 0x001B (DVD+R) Profile : 0x001A (DVD+RW) Profile : 0x0014 (DVD-RW sequential recording) Profile : 0x0013 (DVD-RW restricted overwrite) Profile : 0x0011 (DVD-R sequential recording) Profile : 0x0010 (DVD-ROM) Profile : 0x000A (CD-RW) Profile : 0x0009 (CD-R) (current) Profile : 0x0008 (CD-ROM) > DVD-R and DVD+RW should both be ok. DVD-R can store up to 99 sessions > with several MB wasted between each session. Will aim for DVD+R, DVD-R or DVD+RW and report the result. > If you cannot avoid DVD-RW ... Should be a choice to avoid it. Thanks, ... P. -- mobile: +1 778 951 5147 VoIP: +1 604 670 0140 48.7693 N 123.3053 W
Re: got a mdadm puzzler
On Tuesday, 15 March 2022 15:08:35 EDT Charles Curley wrote: > On Mon, 14 Mar 2022 18:47:52 -0400 > > gene heskett wrote: > > And I am beginning to think in terms of a mobo problem, the bios is > > only seeing the cpu bus a .9 something volts, and the cpu fan at 721 > > rpm, but nothing else and gkrellm isn't even seeing any of that. > > I don't know what CPU voltages are these days (the last I knew, they > were 3.2), but if that is low (see your motherboard manual to be sure), > that could be a power supply problem. Its been that low since this mobo was installed along with a 6 core i5 that has never warmed up its heat sink, even when gkrellm says all cores are flat out at 4.3 ghz, normal, just doing the kmail imap dl stuff is 800 mhz. So, I don't think thats too unusual for this board and cpu combo. And the voltages I have measured, 5 and 12 on a drive connector, are a small amount north of dead on, like the 5 volts is 5.09, and 12 volts is 12.11. So based on gut, plus I'm a CET (altho my billfold card attesting to that is now 50 years old, I sat for that in '72), I think the psu is fine yet. > -- > Does anybody read signatures any more? > > https://charlescurley.com > https://charlescurley.com/blog/ > > . Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis
Re: Archiving content of a directory on a DVD-R.
Hi, pe...@easthope.ca wrote: > Drive type : vendor 'PLEXTOR' product 'DVDR PX-708A' revision '1.06' How old is it ? I find this model mentioned as early as 2003. DVD burning was a novelty back then. > https://www.londondrugs.com/verbatim-dvd-rw---30-pack/L7011505.html Says "You have been blocked". Guessing from the URL: DVD-RW media. I would avoid DVD-RW with an old DVD burner. After DVD+R DL this is the next type of DVD which fails when the burner's sight becomes blurred by age. The internet says that PX-708A does DVD-R, DVD-RW, and DVD+RW. You may ask it by xorriso -outdev /dev/sr0 -list_profiles out A modern DVD burner is supposed to reply something like Profile : 0x0012 (DVD-RAM) Profile : 0x0011 (DVD-R sequential recording) Profile : 0x0015 (DVD-R/DL sequential recording) Profile : 0x0016 (DVD-R/DL layer jump recording) Profile : 0x0014 (DVD-RW sequential recording) Profile : 0x0013 (DVD-RW restricted overwrite) Profile : 0x001A (DVD+RW) Profile : 0x001B (DVD+R) Profile : 0x002B (DVD+R/DL) Profile : 0x0010 (DVD-ROM) Profile : 0x0009 (CD-R) Profile : 0x000A (CD-RW) Profile : 0x0008 (CD-ROM) Profile : 0x0002 (Removable disk) Yours might omit some of those lines. A profile is a set of features. It describes a media role. Most of them are about writing particular media types. Some are about reading. > Not clear which medium is suitable for weekly backups. DVD-R and DVD+RW should both be ok. DVD-R can store up to 99 sessions with several MB wasted between each session. DVD+RW get their sessions emulated by xorriso with no more than 64 MB waste between sessions. In case that the drive announces to support DVD+R: 153 sessions with 4 MB waste between sessions. If you cannot avoid DVD-RW then format them (if profile 0x0013 is announced): xorriso -outdev /dev/sr0 -format as_needed This makes them similar to DVD+RW. DVD+RW and DVD-RW can be re-used. But xorriso will not be willing to overwrite them from scratch unless you explicitely tell it to invalidate the existing ISO 9660 filesystem (or other filesystem formats): xorriso -outdev /dev/sr0 -blank as_needed Have a nice day :) Thomas
Re: Archiving content of a directory on a DVD-R.
From: "Thomas Schmitt" Date: Sat, 26 Feb 2022 18:29:45 +0100 > What messages do you get printed when the CD-or-DVD medium is inserted > and you do: > > xorriso -outdev /dev/sr0 -toc Have this drive in an external case with a blank CD. No DVDs yet. root@joule:/home/root# xorriso -outdev /dev/sr0 -toc xorriso 1.5.2 : RockRidge filesystem manipulator, libburnia project. Drive current: -outdev '/dev/sr0' Media current: CD-R Media status : is blank Media summary: 0 sessions, 0 data blocks, 0 data, 703m free Drive current: -outdev '/dev/sr0' Drive access : exclusive:unrestricted Drive type : vendor 'PLEXTOR' product 'DVDR PX-708A' revision '1.06' Drive id : '148925 ' Media current: CD-R Media product: 97m26s66f/79m59s71f , CMC Magnetics Corporation Media status : is blank Media blocks : 0 readable , 359846 writable , 359846 overall Media summary: 0 sessions, 0 data blocks, 0 data, 703m free Several specifications described here. https://en.wikipedia.org/wiki/DVD_recordable Media which should be available. https://www.londondrugs.com/verbatim-dvd-rw---30-pack/L7011505.html https://www.londondrugs.com/verbatim-dvd%2Br-4.76gb-16x---50-pack---95037/L2167674.html Not clear that this Plextor handles plus and minus. Not clear which medium is suitable for weekly backups. Thanks,... P. -- mobile: +1 778 951 5147 VoIP: +1 604 670 0140 48.7693 N 123.3053 W
Re: got a mdadm puzzler
On Mon, 14 Mar 2022 18:47:52 -0400 gene heskett wrote: > And I am beginning to think in terms of a mobo problem, the bios is > only seeing the cpu bus a .9 something volts, and the cpu fan at 721 > rpm, but nothing else and gkrellm isn't even seeing any of that. I don't know what CPU voltages are these days (the last I knew, they were 3.2), but if that is low (see your motherboard manual to be sure), that could be a power supply problem. -- Does anybody read signatures any more? https://charlescurley.com https://charlescurley.com/blog/
Re: iwd + systemd-networkd + resolvconf wrinkles
On Mon 14 Mar 2022 at 07:15:12 (+), Thomas Pircher wrote: > David Wright wrote: > > I was casting round for a simple way to run iwd + resolvconf + > > systemd-networkd as replacement. > > I run a similar setup, with iwd, systemd-networkd and systemd-resolved. > This has been working without problems on my host for for quite a while > now. As I said, I tried that. > Make a copy of your /etc/resolv.conf file, No point, as there's nothing specific in it, but just what gets sent by DHCP from the router. > then enable and restart the > systemd-resolved service. Finally link the /etc/resolv.conf file to > either /run/systemd/resolve/resolv.conf or > /run/systemd/resolve/stub-resolv.conf. I use the latter: > > # ls -l /etc/resolv.conf lrwxrwxrwx 1 root root 37 Jun 28 2020 > /etc/resolv.conf -> /run/systemd/resolve/stub-resolv.conf (It was created for me.) So, to summarise, I have a three-line /var/lib/iwd/mySSID.psk (Security, PSK, passphrase), a two-line /etc/iwd/main.conf (General, EnableNetworkConfiguration=true), and nothing else: no overrides, no resolvconf package, and no cat5 cable. > You can configure various settings for the DNS resolver in your > systemd-networkd setting and in /etc/systemd/resolved.conf. Like what? > On bookworm you also have the resolvectl tool, which helps debugging DNS > issues. And bullseye has that too. I don't really know how to use it. There seem to be timeouts involved in most cases, so time ping -c 1 foo will typically take 15sec, and host lookups will take 10 or 20sec. That's 10sec, or 20sec, depending on whether the message ;; connection timed out; no servers could be reached is emitted once or twice. I ran resolvectl log-level debug and triedresolvectl query foo on a few addresses. They were even slower, eg: # resolvectl query smtp.lionunicorn.co.uk answered in 57.6 secs. # resolvectl query lionunicorn.co.uk failed with: lionunicorn.co.uk: resolve call failed: Query timed out The debug output is difficult to interpret, though I did notice that it was reporting "cache misses" repeatedly on the same address (but there must be some caching going on, because there was an occasional hit being reported). I also noticed that debug output carries on being emitted after the actual query has finished and returned to a bash prompt; for something like another minute, achieving nothing (repeating a query does it all over again). Everything is comparatively instantaneous when using resolvconf, which is why I chose to continue using it. The idea of "debugging DNS issues" doesn't exactly thrill me. I'm imagining a scenario where I'm sitting in an airport or motel room, having managed to make a connection with iwd and negotiate their captive portal or whatever, and then run into /this/ problem. Cheers, David.
Re: Installing/Preparing Debian on a headless system
On Tue 15 Mar 2022 at 12:41:48 (+), Brian wrote: > On Sun 13 Mar 2022 at 20:05:45 -0500, David Wright wrote: > > On Wed 09 Mar 2022 at 16:18:08 (+), Brian wrote: > > > > > > Sorry, I do not know anything about UEFI and the installer. I will, > > > however, guide you through the steps to use preseeding and get a > > > network console. > > > > That would be most helpful if you would post that. > > Carry out the instructions in sections 1 to 10 at > > https://wiki.debian.org/Installation+Archive+USBStick > > Attached is the relevant portion of my preseed.cfg up to partitioning. > I do that manually. Thanks, very useful. I always install using the network console, so this saves me finding a chair, possibly a screen, and a very boring few minutes. Cheers, David.
Re: packages built with golang
Cousin Stanley wrote: > Cousin Stanley wrote : > > The data is already on your system, so > > there's no transmission happening. > > I do not understand this. ... > Does the Debian package manager > really download package information > for all ~59,000 avaiilabel packages > in anticipation that users will need it > at sometime in the future ? > > That would be suprising to me. Now you are surprised, and informed. apt update acquires all the information about all the packages in the repositories you have configured, and stores that for later use. When you decide to install a package, apt figures out, all on your own machine, what packages are necessary to make your request work, then downloads those packages and installs them. The packages are stored until you decide to delete them, by the way, so you can reinstall them without further network costs. apt list --upgradable doesn't incur a network transaction. If you install apticron, or unattended-upgrades, or any of several other services, you can have your system run apt updates automatically. Some of them can then run package upgrades automatically. Some of them can reboot your system automatically when a kernel upgrade comes along. If you have a flock of machines, one of them can download all the updates for the others so that external network bandwidth is preserved. This can be done on a proxy cache basis (packages are retrieved on demand and held for further requests) or as a complete local mirror of an apt repo. -dsr-
Re: packages built with golang
On 2022-03-15, Cousin Stanley wrote: > >> Whether or not you want to see it >> is a different issue. > > I understand this. > >> The data is already on your system, so >> there's no transmission happening. > > I do not understand this. > > I was under the impression that > package information returned by > > apt-cache show some-package > > for packages that I have not installed > would not be downloaded onto my system > until I actually requested it. 'apt update' downloads the package metadata; when if you give apt the update flag you request exactly that.
Re: packages built with golang
On Tue, Mar 15 2022 at 08:28:41 AM, Cousin Stanley wrote: > > > > I was under the impression that > package information returned by > > apt-cache show some-package > > for packages that I have not installed > would not be downloaded onto my system > until I actually requested it. > > If that is the case then a transmission > would be happening when one made > such a request. > > Does the Debian package manager > really download package information > for all ~59,000 avaiilabel packages > in anticipation that users will need it > at sometime in the future ? > Yes, this is what happens. See the files in the /var/lib/apt/lists directory. > That would be suprising to me. > See the description of the update operation in the apt-get manpage. > -- regards, kushal
Re: packages built with golang
On Tue, Mar 15, 2022 at 08:28:41AM -0700, Cousin Stanley wrote: > I was under the impression that > package information returned by > > apt-cache show some-package > > for packages that I have not installed > would not be downloaded onto my system > until I actually requested it. You are mistaken. All of the information given by apt-cache is retrieved from the package metadata that is CACHED on your system already. That's the whole point of the command's name. The files it uses are in /var/lib/apt/lists/ if you want to look at them. On my system: unicorn:~$ grep -A 10 'Package: hugo' /var/lib/apt/lists/ftp.us.debian.org_debian_dists_bullseye_main_binary-amd64_Packages Package: hugo Source: hugo (0.80.0-6) Version: 0.80.0-6+b5 Installed-Size: 47774 Maintainer: Debian Go Packaging Team Architecture: amd64 Depends: libsass1 (>= 3.6.3), libc6 (>= 2.4) Recommends: git Description: Fast and flexible Static Site Generator written in Go Homepage: https://gohugo.io/ Built-Using: go-md2man-v2 (= 2.0.0+ds-5), golang-1.15 (= 1.15.9-6), golang-blackfriday (= 1.6.0-1), [...] Yours may have a different filename, depending on which mirrors you use.
Re: packages built with golang
Cousin Stanley wrote : >> What I don't understand is the necessity >> to transmit a string of 3933 bytes >> for 87 golang packages for example >> when a link to the same would suffice >> for those that actually require it. Dan Ritter wrote : > The package manager needs it. I understand this. > Whether or not you want to see it > is a different issue. I understand this. > The data is already on your system, so > there's no transmission happening. I do not understand this. I was under the impression that package information returned by apt-cache show some-package for packages that I have not installed would not be downloaded onto my system until I actually requested it. If that is the case then a transmission would be happening when one made such a request. Does the Debian package manager really download package information for all ~59,000 avaiilabel packages in anticipation that users will need it at sometime in the future ? That would be suprising to me. On the Debian Bullseye system that I'm currently using I only have a fraction of the available packages installed $ dpkg -l | wc -l 2366 > > You have options. Greg provided you with one. I understand this and the use of grep -v to supress output and fully appreciate the suggestion. However, in such cases as the long Built-Using examples, the so-called information overload has already ocured which might have been saved if a link had been provided instead. -- Stanley C. Kitching Human Being Phoenix, Arizona
Re: packages built with golang
On Tue, Mar 15, 2022 at 06:36:52AM -0700, Cousin Stanley wrote: > to...@tuxteam.de wrote: > > > No, it seems you haven't understood. > > > > I assure you that I do understand the need > for access to the Built-Using list. > > What I don't understand is the necessity > to transmit a string of 3933 bytes > for 87 golang packages [...] You conveniently snipped out my attempt at explaining that. As a sysadmin I sure want to see that info. We all got by now that you don't. Now selecting a default in this situation can go in this or that direction. Cheers -- t signature.asc Description: PGP signature
Re: Installing/Preparing Debian on a headless system
On 2022-03-15 12:41, Brian wrote: Attached is the relevant portion of my preseed.cfg up to partitioning. I do that manually. "d-i netcfg/wireless_essid string MI5_Listening_Station_102" =O) -- Key ID4BFEBB31
Re: packages built with golang
Cousin Stanley wrote: > to...@tuxteam.de wrote: > > > No, it seems you haven't understood. > > > > I assure you that I do understand the need > for access to the Built-Using list. > > What I don't understand is the necessity > to transmit a string of 3933 bytes > for 87 golang packages for example > when a link to the same would suffice > for those that actually require it. The package manager needs it. Whether or not you want to see it is a different issue. The data is already on your system, so there's no transmission happening. > As a casual user only perusing package information > for an introductory cursory look, > a list entailing that level of detail > doesn't seem to be necessary. You have options. Greg provided you with one. -dsr-
Re: packages built with golang
to...@tuxteam.de wrote: > No, it seems you haven't understood. > I assure you that I do understand the need for access to the Built-Using list. What I don't understand is the necessity to transmit a string of 3933 bytes for 87 golang packages for example when a link to the same would suffice for those that actually require it. As a casual user only perusing package information for an introductory cursory look, a list entailing that level of detail doesn't seem to be necessary. -- Stanley C. Kitching Human Being Phoenix, Arizona
Re: Disques ssd sensibles aux coupures brutales ?
Le sam. 12 mars 2022 à 00:30, Daniel Caillibaud a écrit : > > Bonjour, > > Je reviens sur ce thread parce que : > > Le 04/03/22 à 17:21, Olivier a écrit : > > Mon objectif est d'éviter d'endommager un disque (toujours de type SSD ou > > NVMe) à cause d'une > > coupure brutale de courant. > > m'étonne un peu. > > Un disque ssd est vraiment sensible à une coupure brutale ? > > Je me souviens™ du devoir de parquer les disques avant d'éteindre un PC, mais > c’était au siècle > dernier ! > > La question est vraiment très intéressante. Merci beaucoup de la poser. Au 21ème siècle, j'ai eu de multiples pannes de machines refusant de démarrer à cause d'une incohérence dans le système de fichier. La solution était souvent de déclencher un simple fsck. C'est une opération assez difficile sur des machines headless sans aucun informaticien aux alentours. Par contre, si ma mémoire ne me trompe pas, je crois n'avoir rencontré ce cas qu'avec des disques SATA. Je sais que j'ai eu des coupures électriques sauvages et sans conséquence avec des disques NVMe mais je ne sais pas si on peut attribuer l'absence de conséquence à la chance ou à autre chose. Néanmoins, je crois que des disques NVMe ont des mécanismes qui les protègent contre les coupures sauvages. Couplés à des onduleurs télé-administrables et des cartes mères supportant la fonction "Power ON: Last State", on a peut-être une solution simple re-démarrant automatiquement même si la route est longue (cf [1]) [1] https://www.tomshardware.com/news/sk-hynix-sabrent-rocket-ssds-data-loss
Re: Installing/Preparing Debian on a headless system
On Sun 13 Mar 2022 at 20:05:45 -0500, David Wright wrote: > On Wed 09 Mar 2022 at 16:18:08 (+), Brian wrote: > > > > Sorry, I do not know anything about UEFI and the installer. I will, > > however, guide you through the steps to use preseeding and get a > > network console. > > That would be most helpful if you would post that. Carry out the instructions in sections 1 to 10 at https://wiki.debian.org/Installation+Archive+USBStick Attached is the relevant portion of my preseed.cfg up to partitioning. I do that manually. -- Brian. # This file is designed to be used with 'priority=critical'. (The # priority would be 'high' without this entry). # (https://www.debian.org/releases/stable/i386/ch05s03.html.en). debconf debconf/priority string critical # Provide non-free firmware files directly to the installer. It now # does not need to search for them in other packages. d-i preseed/early_command string\ cp -a /hd-media/files/firmware /lib # If possible, netcfg will choose interfaces that have links. This # makes it skip displaying a list if there is more than one interface. # The first connected network interface will be selected and used. # # The IP address may be changed after first boot. d-i netcfg/choose_interface select auto d-i netcfg/disable_autoconfig boolean true d-i netcfg/get_ipaddress string 192.168.7.99 d-i netcfg/get_netmask string 255.255.255.0 d-i netcfg/get_gateway string 192.168.7.1 d-i netcfg/get_nameservers string 192.168.7.1 d-i netcfg/confirm_static boolean true # This sets the hostname for the system; it ignores names provided by # DHCP or DNS. 'hostname=' on the command line of the installer does not # override it. Change the hostname after installation. d-i netcfg/hostname string bullseye # Use ifupdown irrespective of whether a DE (which would bring in # network manager) is installed or whether the connection is wired or # wireless. d-i netcfg/target_network_config select ifupdown # Wireless stuff, if it is available. d-i netcfg/wireless_show_essids select manual d-i netcfg/wireless_essid string MI5_Listening_Station_102 d-i netcfg/wireless_security_type select wpa d-i netcfg/wireless_wpa string xxx # Users and passwords. d-i passwd/user-fullname string x yy d-i passwd/username string brian d-i passwd/root-password password PASWORD d-i passwd/root-password-again password PASSWORD d-i passwd/user-password password PASSWORD d-i passwd/user-password-again password PASSWORD # Log in from another machine on the local network. Obtain the # IP of the machine booted with d-i from 'arp-scan' if there # isn't a way to use 'ip a' (no monitor attached) and a static # address has not been preeseeded. Then do # # ssh installer@IP_ADDRESS # or # ssh -o GlobalKnownHostsFile=/dev/null -o UserKnownHostsFile=/dev/null installer@IP_ADDRESS # # and give the password. Change the Debconf priotity back to # "critical" to continue with the remainder of the directives # in preseed.cfg. # # To change console precede the console number with Ctrl-A. For # example, Ctrl-A 4. The console has screen support. # (https://lists.debian.org/debian-boot/2016/02/msg00547.html). d-i anna/choose_modules string network-console d-i network-console/password password PASSWORD d-i network-console/password-again password PASSWORD # network-console becomes available when a network connection is # established. This line ensures preseeding continues up to the # partitioning stage, which is done manually. d-i network-console/start select continue
Re: disque plein mais pas plein
Bonjour, Bon en fait j'utilise zoneminder, et c'est lui qui m'a rempli mes inodes. Merci
Re: disque plein mais pas plein
Salut, J'ai eu le problème il y a peu (mon .bash_history s'en souvient) et j'ai dû refaire une partition n'ayant pas trouvé comment utiliser tune2fs pour modifier le nombre d'inodes. L'option de mkfs.ext4 est -N pour modifier ce nombre. Pour recopier le contenu d'une partition aisément sans s'occuper de FS monté dans des répertoires de la partition à copier, tu peux monter une seconde fois cette partition dans un autre répertoire puis avec : find . | cpio -pd /path/to/newFS Avec cette commande lancée à l'endroit où est montée la partition à copier, tu recopies l'intégralité du FS et uniquement ce FS. Le jeu. 17 févr. 2022 à 10:08, Luc Novales a écrit : > Bonjour, > Le 16/02/2022 à 16:28, Benoît SZCZYGIEL a écrit : > > Bonne réponse collective, effectivement les inodes sont à 100%. Existe t-il > une solution pour les augmenter sans réformateur? Je n'ai pas trouvé. > Merci > > Effacer quelques fichiers pour pouvoir investiguer serait peut être une > solution. Ensuite, augmenter le nombre d'inodes sans savoir ce qui les > consomme ne va que retarder un prochain problème. > > A part un logiciel crée beaucoup de petits fichiers comme du cache http, > le nombre d'inodes par défaut est suffisant. > > La commande : sudo du --inodes -s /var/* > > peut aider à diagnostiquer. > > Bonne journée, > > Luc. >
Re: Comment éteindre un serveur proprement pour permettre le redémarrage automatique ?
Le 2022-03-15 10:16, Mathias Dufresne a écrit : Une option peut être le wake-on-lan si le problème du redémarrage automatique ne fonctionne pas. Malheureusement ça nécessite une machine allumée donc soit par un démarrage manuelle, soit une machine qui arrive a se réveiller toute seule après retour de l'électricité. P't'être même qu'un vieux PC qui s'allume quand le courant revient pour lancer des paquets wake-on-lan et qui s'éteigne une fois sa tâche terminée pourrait faire l'affaire : à la fin de la prochaine coupure, cette machine devrait se réveiller et lancer les paquets... Sur mon onduleur, j'ai : * des prises secourues et qui offrent une protection contre la foudre ; * des prises *non* secourues, qui offrent une protection contre la foudre. Je pense qu'en branchant sur une prise *non* secourue une carte minimaliste, du genre Raspberry Pi Zero W : https://www.kubii.fr/cartes-raspberry-pi/1851-raspberry-pi-zero-w-kubii-3272496006997.html On peut aisément – et à moindre cout – fournir le service attendu. Si le wifi n'est pas disponible, il faudra opter pour un modèle plus « luxueux » (au regard de l'usage qui en est fait), disposant d'une interface Ethernet, du genre Raspberry Pi 3 : https://www.kubii.fr/home/1628-raspberry-pi-3-modele-b-1-gb-kubii-5060214370264.html Il semblerait que l'on puisse même faire cela avec un simple ESP32 : https://github.com/mkttanabe/ESP32_WakeOnLan Sébastien -- Sébastien Dinot Ne goûtez pas au logiciel libre, vous ne pourriez plus vous en passer ! https://www.palabritudes.net/
Re: Disques ssd sensibles aux coupures brutales ?
Salut, @home, je ne m'inquiète absolument pas de ça. Sans onduleur, mon mini-serveur qui fait office de NAS au passage redémarre correctement depuis des années après chaque coupure d'électricité. Juste pour te rassurer ^^ Je dirai que le risque est lié à l'utilisation des disques. Pour qu'un disque présente un problème après une coupure, soit c'est mal géré par le système (au sens très large) soit c'est que le disque n'a pas eu le temps de finir le travail d'écriture en cours. Or les SSD écrivent plutôt vite ce qui limite les risques d'une écriture non terminée. Le risque augmente avec l'utilisation du FS mais tu parles de NAS perso, ça sert le plus souvent à la lecture et comme il est "perso", le nombre d'écriture doit être relativement restreint. Je suis par contre assez surpris de lire que tu as besoin de fsck régulièrement vu que (selon moi, c'est pas une vérité gravée dans le marbre ; ) le problème est forcément lié à l'écriture et que les systèmes n'écrivent pas grand chose hormis leurs logs, peut-être vérifier sur quelle partition / FS s'appliquent ces fsck, voire si ce n'est pas encore le cas, séparer les logs du reste du systèmes (en terme de partitions) afin que le risque soit cloisonné... Le sam. 12 mars 2022 à 00:30, Daniel Caillibaud a écrit : > Bonjour, > > Je reviens sur ce thread parce que : > > Le 04/03/22 à 17:21, Olivier a écrit : > > Mon objectif est d'éviter d'endommager un disque (toujours de type SSD > ou NVMe) à cause d'une > > coupure brutale de courant. > > m'étonne un peu. > > Un disque ssd est vraiment sensible à une coupure brutale ? > > Je me souviens™ du devoir de parquer les disques avant d'éteindre un PC, > mais c’était au siècle > dernier ! > > Assez rapidement les constructeurs ont ajoutés du park auto à l’extinction > (j’imagine le nb de > plaintes qu’ils ont eu de gens ayant dépensé des fortunes pour un hd, > parti en fumée parce que > qqun avait déplacé un PC éteint), puis du park auto à la coupure de > courant (ils ont réinventé > le ressort). > > Depuis les ssd y’a plus de tête risquant de se retrouver au mauvais > endroit au mauvais moment, > donc en cas de coupure de courant je veux bien qu’il y ait un risque sur > les datas (et encore, > avec les fs journalisés ça devrait plus trop être le cas), mais un risque > sur le matériel ? > > C’est toujours d’actualité ? > > Si oui faut vraiment que je m’inquiète car sur mon PC actuel j’ai du > reboot hard ~3×/semaine > depuis 1an 1/2, du "recovered inode" à chaque reboot après un plantage, > mais heureusement les > disques sont toujours là (un nvme et un disque HD classique à plateaux, > moins sollicité). > > Rien d’ironique, je suis une buse en hardware et peux très bien avoir de > fausses idées reçues, > si qqun qui sait peut confirmer / infirmer ça m’intéresse. > > > PS: ça ne remet pas en cause l’intérêt d’un onduleur, mais pour du NAS > perso, ça me paraît un > peu overkill (je ne parle pas du coût environnemental, changer les > batteries tous les 2ans, > toussa, juste du coût humain pour s’occuper de l’onduleur et sa > communication avec la machine, > plutôt que de laisser la machine redémarrer toute seule quand le courant > revient). > > -- > Daniel > > Lorsque j'ai été kidnappé, ma mère a réagi tout de suite: elle a sous-loué > ma chambre. > Woody Allen > >
Re: Comment éteindre un serveur proprement pour permettre le redémarrage automatique ?
Salut, L'option mentionnée du BIOS pour le redémarrage automatique ne fonctionne que si l'alimentation électrique a été coupée. Ici, sur une carte mère sans doute très différente de la tienne, après un poweroff suivi d'une coupure électrique, la machine redémarre automatiquement lorsque l'électricité est à nouveau disponible... à condition d'attendre suffisamment pour que les condensateurs se vident et que la carte comprenne que l'électricité est coupée. Une option peut être le wake-on-lan si le problème du redémarrage automatique ne fonctionne pas. Malheureusement ça nécessite une machine allumée donc soit par un démarrage manuelle, soit une machine qui arrive a se réveiller toute seule après retour de l'électricité. P't'être même qu'un vieux PC qui s'allume quand le courant revient pour lancer des paquets wake-on-lan et qui s'éteigne une fois sa tâche terminée pourrait faire l'affaire : à la fin de la prochaine coupure, cette machine devrait se réveiller et lancer les paquets... En espérant que ça puisse t'être utile ; ) Le ven. 4 mars 2022 à 23:06, NoSpam a écrit : > Comme déjà proposé, nut permet la gestion des onduleurs en USB, série ou > ethernet. > > Le 04/03/2022 à 17:21, Olivier a écrit : > > Bonjour, > > > > J'envisage de protéger des serveurs distants (de type NUC) avec un > > "onduleur administrable". Mon objectif est d'éviter d'endommager un > > disque (toujours de type SSD ou NVMe) à cause d'une coupure brutale de > > courant. > > > > J'accepte que les services soient interrompus tant que dure la panne > > de courant mais j'aimerai qu'idéalement, les services redémarrent sans > > intervention humaine quand le courant revient (si par chance, celui-ci > > devait revenir sans action humaine). > > > > > > Imaginons qu'un serveur distant protégé par cet onduleur > > administrable, reçoive de ce dernier ou d'ailleurs, la notification > > d'une panne de courant prolongée. > > > > Quelle commande d'extinction-hibernation doit-il émettre afin : > > 1- qu'il consomme le minimum d'énergie tant que dure la panne de courant > > 2- qu'il re-démarre dès que le courant revient. > > > > J'ai vu dans le BIOS une option "After Power Failure: Stay Off/Power > > On Normal Boot/Power On PXE". Je l'ai essayé mais elle ne fonctionne > > après une commande poweroff, ce qui me semble logique. > > > > Une idée ? > > > > Slts > >
Re : Re: Régler luminosité écran carte vidéo Intel
On 15/03/2022 02:02, k6dedi...@free.fr wrote: Par contre, les touches clavier du portable, F11 (luminosité -) et F12 (luminosité +), bien pratiques, ne fonctionnent pas du tout et pourquoi ? Parce qu'il faut un programme spécifique pour gérer ça. Fais une recherche dans Synaptic sur la marque de ton portable pour voir s'il y un paquet qui gère les touches spéciales.
Re: voltage monitoring Q
On Tuesday, 15 March 2022 02:43:59 EDT Richard Hector wrote: > On 13/03/22 21:15, gene heskett wrote: > > they are the last seacrate drives I'll own... Ever. > > Lots of brands seem to go through bad patches. Even just bad batches. > > For stuff I care about, I use RAID1 (mdraid), on NAS drives, from mixed > manufacturers. So I'll have a pair consisting of a Seagate IronWolf > and a WD Red. That way, if one dies, it's less likely that the other > will immediately follow suit. > > Except if I'm buying a 'server' from the local shop - they insist that > mixing drives will be less reliable, but don't give reasons or > evidence. I insisted for my home backup server, with a Supermicro > server board, but they put 'Custom Workstation' on the invoice - I > don't think they were prepared to call it a server. > > Richard I like that way of thinking Richard. For long term data retention it sounds more reliable to get rid of the cloned twins with the same disease problem. Thanks. Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis
Re: voltage monitoring Q
On 13/03/22 21:15, gene heskett wrote: they are the last seacrate drives I'll own... Ever. Lots of brands seem to go through bad patches. Even just bad batches. For stuff I care about, I use RAID1 (mdraid), on NAS drives, from mixed manufacturers. So I'll have a pair consisting of a Seagate IronWolf and a WD Red. That way, if one dies, it's less likely that the other will immediately follow suit. Except if I'm buying a 'server' from the local shop - they insist that mixing drives will be less reliable, but don't give reasons or evidence. I insisted for my home backup server, with a Supermicro server board, but they put 'Custom Workstation' on the invoice - I don't think they were prepared to call it a server. Richard
Re: packages built with golang
On Mon, Mar 14, 2022 at 05:23:02PM -0700, Cousin Stanley wrote: > Dan Ritter wrote: > > > > > Not having Built-Using is just like not having dependencies. > > Thanks for the explanation. > > I can understand the need for the Built-Using list > for the developers that need it. No, it seems you haven't understood. Assume some security issue becomes known for, say, libssl. The developers fix it. The Debian developer integrates the fix into the package (if necessary, by backporting. You, the user, do an apt-get upgrade. The fixed version gets installed in your system, every program using that library is now fine. With go, you have, say, some crypto library linked STATICALLY to 15 programs from as many packages. The bug gets fixed and... nothing happens. Unless... your system knows that those 15 packages were "Built-Using:". That's why you need that metadata. As for why they are shown to you, the system administrator -- well, the installer's decision to install those 15 packages has to be transparent to you. It is, after all, the dark sister of "Depends:" in the statical linking world. Cheers -- t signature.asc Description: PGP signature
