Table des matières
4 billet(s) pour janvier 2026
| AWX sur K8S Kind - partage de fichier pour les blob - Execution pods | 2026/01/26 10:15 | Jean-Baptiste |
| Notes rsh rcp | 2026/01/21 18:08 | Jean-Baptiste |
| Git - Duplication d'un dépôt | 2026/01/19 10:22 | Jean-Baptiste |
| Exemple simple de conf Nagios | 2026/01/14 10:07 | Jean-Baptiste |
Notes TCP/IP Jumbo Frame Trame géante MTU
Voir :
Peut être utile pour du SAN / NAS :
Connaître la MTU d'une interface
netstat -i ip link |grep mtu
Changer provisoirement la MTU d'une inferface
#ifconfig eth1 mtu 9000 up ip link set eth1 mtu 9000
Changer la MTU sur RedHat
/etc/sysconfig/network-scripts/ifcfg-enp0s9
HWADDR=08:00:27:D9:8B:E7 TYPE=Ethernet BOOTPROTO=none IPADDR=192.168.25.21 PREFIX=24 DEFROUTE=yes IPV4_FAILURE_FATAL=yes IPV6INIT=no NAME=enp0s9 UUID=5f82629e-5117-4d41-902d-090d82053cb2 ONBOOT=yes MTU=9000
Changer la MTU sous RedHat avec NetworkManager
nmcli connection modify eno2 802-3-ethernet.mtu 9000 nmcli connection show eno2 |grep mtu
Changer la MTU d'une interface sur Debian
/etc/network/interfaces.d/eth1
auto eth1
iface eth1 inet static
address 192.168.56.12
netmask 255.255.255.0
mtu 9000
Appliquer les changements
systemctl restart networking
Connaitre la MTU entre deux points
# traceroute --mtu 192.168.56.1 traceroute to 192.168.56.1 (192.168.56.1), 30 hops max, 65000 byte packets 1 192.168.56.1 (192.168.56.1) 0.223 ms F=1500 * 0.446 ms
F=1500
Tester
1472 = 1500 - 28
Voir aussi /proc/sys/net/ipv4/ip_no_pmtu_disc
# ping -M do -s 1472 192.168.6.130 PING 192.168.6.130 (192.168.6.130) 1472(1500) bytes of data. 1480 bytes from 192.168.6.130: icmp_req=1 ttl=255 time=0.095 ms 1480 bytes from 192.168.6.130: icmp_req=2 ttl=255 time=0.127 ms ^C --- 192.168.6.130 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 999ms rtt min/avg/max/mdev = 0.095/0.111/0.127/0.016 ms # ping -M do -s 1473 192.168.6.130 PING 192.168.6.130 (192.168.6.130) 1473(1501) bytes of data. ^C --- 192.168.6.130 ping statistics --- 4 packets transmitted, 0 received, 100% packet loss, time 2999ms
Path MTUs are recorded in the routing cache. This can interfere with testing, and in particular, can make a remote MTU restriction appear to be a local one. You can clear the cache using the ip route command.
ip route flush cache
Cherche rapide de fichier grâce à updatedb et locate
Voir aussi :
- plocate (a vocation de remplacer le bon vieux mlocate)
Exemple 1
# systemd-run -p IOWeight=10 updatedb updatedb -l 0 -o db_file -U source_directory locate -d db_file NOM_FICHIER
Exemple 2
updatedb -l 0 --prunenames ".git .bzr .hg .svn .gz" --prunepaths "${HOME}/tmp ${HOME}/mnt" -U ~ -o ~/mlocate.db locate -d ~/mlocate.db NOM_FICHIER locate -d ~/mlocate.db NOM_FICHIER -0 | xargs -0 ls -ld
Note : prunepaths doit contenir des chemin absolus
http://fr.wikipedia.org/wiki/Locate http://www.thegeekstuff.com/2012/03/locate-command-examples/
Examples To create a private mlocate database as an user other than root, run
updatedb -l 0 -o db_file -U source_directory
Source : http://linux.die.net/man/8/updatedb
Notes TCL
tclsh
Equivalent de bash -c
runner.tcl
#!/usr/bin/tclsh eval $argv
tclsh runner.tcl puts hello
Notes Tahoe-LAFS
https://www.digitalocean.com/community/tutorials/tahoe-lafs
Brouillon. (source http://killyourtv.i2p/tahoe-lafs/running/)
If you're not using a multi-introducer capable Tahoe installation, most are using killyourtv's introducer to connect to the grid so that would be a good choice to use as a primary introducer. Updating your Tahoe-LAFS installation is recommended. If you're using the multi-introducer build, copy and paste the following into ~/.tahoe/introducers:
apt-get install tahoe-lafs
tahoe --version
Warning: dependency 'cryptography' (version '0.6.1') found by pkg_resources not found by import. Warning: dependency 'six' (version '1.8.0') found by pkg_resources not found by import. Warning: dependency 'cffi' (version '0.8.6') found by pkg_resources not found by import. Warning: dependency 'pycparser' (version '2.10') found by pkg_resources not found by import. For debugging purposes, the PYTHONPATH was None install_requires was ['setuptools >= 0.6c6', 'zfec >= 1.1.0', 'simplejson >= 1.4', 'zope.interface == 3.6.0, == 3.6.1, == 3.6.2, >= 3.6.5', 'Twisted >= 11.0.0', 'foolscap >= 0.6.3', 'pyOpenSSL', 'Nevow >= 0.6.0', 'pycrypto == 2.1.0, == 2.3, >= 2.4.1', 'pyasn1 >= 0.0.8a', 'mock >= 0.8.0', 'pycryptopp >= 0.6.0'] sys.path after importing pkg_resources was /usr/bin: /usr/lib/python2.7: /usr/lib/python2.7/plat-x86_64-linux-gnu: /usr/lib/python2.7/lib-tk: /usr/lib/python2.7/lib-old: /usr/lib/python2.7/lib-dynload: /usr/local/lib/python2.7/dist-packages: /usr/lib/python2.7/dist-packages: /usr/lib/pymodules/python2.7
Voilà ce qui se passe quand on install des paquets de Debian Stretch sur Debian Jessie. C'est moche. Vive les virtualenv
Activer les dêpots de Debian Stretch
apt-get install tahoe-lafs python-cryptography python-pyasn1 python-cffi python-ply python-pycparser python-numpy
Autre solution
virtualenv tahoenv cd tahoenv source bin/activate pip install --upgrade allmydata-tahoe
tahoe create-node
Node created in '/home/exploit/.tahoe' Please set [client]introducer.furl= in tahoe.cfg! The node cannot connect to a grid without it. Please set [node]nickname= in tahoe.cfg
Client node
A client node is used to communicate with storage nodes in order to store and retrieve files on the grid. There is a web interface (default on http://127.0.0.1:3456/) and a command line interface (via tahoe) for accessing this data. Other interfaces are available, see the Tahoe-LAFS site for details.
tahoe --version
tahoe create-client
Specify a nickname for your node. This is shown to all nodes in the grid so choose the nickname carefully. Edit the configuration file ~/.tahoe/tahoe.cfg. Modify the following options, the other values should be left at the defaults for now:
[node] nickname = yourclientnodename web.port = tcp:3456:interface=127.0.0.1 http_proxy = 127.0.0.1:4444 tub.location =
~/.tahoe/tahoe.cfg
uncomment tub.location otherwise your (at worse external!) IP will be leaked to introducers: In client section remove None word in introducer.furl,
Uncomment tub.location but leave it empty. This will prevent your other interfaces besides 127.0.0.1 from being broadcast to the introducer. /!\ WARNING: Make sure that you do this to avoid leaking your IP address! /!\
[node] nickname = YourRandomNickname web.port = tcp:3456:interface=127.0.0.1 web.static = public_html http_proxy = 127.0.0.1:4444
- IMPORTANT
- You NEED to configure tub.location when running
- an I2P-enabled Tahoe-LAFS node. Otherwise your IP
- can be leaked to the introducers!
- Read the docs and/or come to #tahoe-lafs on Irc2P for assistence.
- tub.port =
tub.location =
- log_gatherer.furl =
- timeout.keepalive =
- timeout.disconnect =
- ssh.port = 8022
- ssh.authorized_keys_file = ~/.ssh/authorized_keys
[client]
- Which services should this client connect to?
introducer.furl = helper.furl =
- key_generator.furl =
- stats_gatherer.furl =
Create introducers file (without .txt extension) and paste the list of introducers from http://killyourtv.i2p/tahoe-lafs/introducers/
Upload
tahoe create-alias YOURALIAS
Upload your website into grid using tahoe cp command:
tahoe cp -r -v “C:\pathtoyourdeepsite” YOURALIAS
This is very time consuming process, uploading a single 50MB files could be faster than uploading 150 files of 5MB total, so consider for example converting all small images into base64 and embedding them into html file.
After uploading, check if all files are uploaded:
tahoe ls YOURALIAS
Type
tahoe list-aliases
or open .tahoe\private\aliases file to get your private keys, and go into website: http://127.0.0.1:3456/uri/URI:DIR2:<yourprivatekey>.
Click More info on this directory and copy your public key from Directory readcap, so your website address is: http://127.0.0.1:3456/uri/URI:DIR2-RO:<yourpublickey>/index.htm.
Never share address without URI:DIR2-RO (read only) otherwise others can vandalize/delete your site! Older files will be deleted from the network (grid). To prevent this, run the following command from time to time (once per 2 months):
tahoe deep-check -v --repair --add-lease YOURALIAS
Introducer
Setting up and running your own introducer on I2P involves a few simple steps. The first of these steps is to create the introducer node: $ tahoe create-introducer ~/.tahoe/introducer The introducer can exist anywhere on your filesystem. If you don't want it to be contained within your ~/.tahoe directory, simply specify a different directory. Next, edit ~/.tahoe/introducer/tahoe.cfg. Fill out the web.port. In this example we'll use 3460: [node] web.port = tcp:3460:interface=127.0.0.1 web.static = public_html Start your introducer: $ tahoe start ~/.tahoe/introducer $ cat introducer.port 17337 Choose Type: Standard (don't pick HTTP; it won't work). Point it to the port listed in introducer.port (in the example above, 17337). Start the server tunnel and take note of the Base32 destination. Edit ~/.tahoe/introducer/tahoe.cfg once more and set the tub.location to the introducer's Base32 destination. Restart your introducer node once more. $ tahoe restart ~/.tahoe/introducer At this point you have an introducer running and a tunnel pointing to it so it can be used on I2P. Of course, no one will be able to use your introducer without its existence being known. You will find the introducer.furl to give to others in the file ~/.tahoe/introducer/private/introducer.furl $ cat ~/.tahoe/introducer/private/introducer.furl pb://c6w5ernw7y7rp3uwmdyu5clujyt2y4m4@w2zrwz5gplkkufix7cb4gmxfbrkwg2abnsgk62bm5iifzlahe7kq.b32.i2p/introducer
Introducers on Tahoe-LAFS define the grid and their addresses should be published to everyone who wishes to connect to this grid. The example below points to a test introducer operated by KillYourTV. To use KYTV's introducer, add the following introducer.furl to tahoe.cfg. [client] introducer.furl = pb://c6w5ernw7y7rp3uwmdyu5clujyt2y4m4@w2zrwz5gplkkufix7cb4gmxfbrkwg2abnsgk62bm5iifzlahe7kq.b32.i2p/introducer Add each introducer that you want to use to ~/.tahoe/introducers -- one per line! For example, I currently have $ cat ~/.tahoe/introducers pb://c6w5ernw7y7rp3uwmdyu5clujyt2y4m4@w2zrwz5gplkkufix7cb4gmxfbrkwg2abnsgk62bm5iifzlahe7kq.b32.i2p/introducer pb://exupps5kk3amc5iq4q6f5ahggkm4s5fl@oj7cffq5fnk46iw3i3h2sdgncxrqbxm7wh6i4h2cbpmqsydygkcq.b32.i2p/introducer pb://md2tltfmdjvzptg4mznha5zktaxatpmz@5nrsgknvztikjxnpvidlokquojjlsudf7xlnrnyobj7e7trdmuta.b32.i2p/introducer pb://fmcbgy7zd6ubrbphilmrlocvb7f327z5@gdr3tt5uewgnm7r7xn54k2qikf2kuwwegjjsnkz44pjticcacsua.b32.i2p/introducer pb://tq7rx35yopkvodmsxkqra4qqkbho3yaa@6ga2r2h2fyq6tzcyh6bf3hpio3i7r4edadbq7l4wnh4y62taj6ia.b32.i2p/introducer pb://cys5w43lvx3oi5lbgk6liet6rbguekuo@sagljtwlctcoktizkmyv3nyjsuygty6tpkn5riwxlruh3f2oze2q.b32.i2p/introducer pb://r3bs6joub24gtsofe7ohnnjcnwfmo2jy@qaihdh5z7osn7tc3326ahv3z46badiuaulff43wchmap7skg7euq.b32.i2p/42mrbm7zxmjemz6hzejo3i7aunx4eoun You will need to restart your tahoe node after editing ~/.tahoe/introducers with tahoe restart. Any existing introducer.furl entries in tahoe.cfg will be automatically copied to ~/.tahoe/introducers.
Node Operation
To start your node as a daemon:
$ tahoe start
You can now access your node via the web interface at http://127.0.0.1:3456.
To stop your node if running as daemon:
$ tahoe stop
Storage node
~/.tahoe/tahoe.cfg
[node] tub.port = tcp:3459:interface=127.0.0.1 tub.location = n6k7tu37diq4obmcmv3f144oxxxnehp6njz5lwu4hllp5gp6hbga.b32.i2p
[storage] enabled = true
Optionally you may define how much disk space is reserved with reserved_space:
reserved_space = 5GB
Notes SystemD
Voir :
Hardening Security focused systemd configuration
Lister les services
systemctl list-unit-files systemctl list-units
Lister les services inactifs
systemctl list-units -all --state=inactive
Lister les services KO
systemctl -a |grep dead
Lister les services actifs
# SystemD systemctl list-units --type service --state active # Avant SystemD chkconfig --list |grep -E '[1-5]:on'
Savoir si le service est activé ou désactivé
systemctl is-active httpd systemctl is-enabled httpd
Relire la conf
systemctl daemon-reload
Les différents types de services
Runlevel et Target
| Runlevel | Target |
| 0 | poweroff.target |
| 1 | rescue.target |
| 2, 3, 4 | multi-user.target |
| 5 | graphical.target |
| 6 | reboot.target |
Contrôler des services (Units) SystemD distants avec Hirte
Voir :
Exemple
hirtectl start rpi4 httpd.service
Debug
systemctl show --property=Environment docker
Autres
If you don't like systemctl's auto-paging feature
export SYSTEMD_PAGER=
Logs entre deux date
journalctl --since "2019-10-16 06:00" --until "2019-10-16 10:00"
Disable systemd redirection
export _SYSTEMCTL_SKIP_REDIRECT=1
tmpfile
[Service] PrivateTmp=true
—
man -k tmpfiles
/etc/tmpfiles.d
- /etc/tmpfiles.d/python.conf
x /tmp/pymp-*
reboot
systemctl reboot
Lang Locale
Voir Notes langue lang locale - SystemD
#timedatectl list-timezones timedatectl set-timezone Europe/Paris
Pb
RemoveIPC - Quand Poettering fait des siennes
Problème si compte UID > 1000 à cause de RemoveIPC
https://linuxfr.org/users/neologix/journaux/systemd-attention-a-removeipc
RemoveIPC Directive
A new option called RemoveIPC was introduced in RHEL 7.2 through Systemd v219. When set to yes, this option forces a cleanup of all allocated inter-process communication (IPC) resources linked to a user leaving his last session. If a daemon is running as a user with a uid number >=1000, it may crash.
This option should always be set to no by default but, due to the logic of package upgrade, it is highly advisable to set RemoveIPC=no in the /etc/systemd/logind.conf file followed by # systemctl restart systemd-logind (source).
Source : https://www.certdepot.net/rhel7-changes-between-versions/
