ich wollte die linbo-syncronisation via lm-web testen und ein PC einfach mal via wol starten.
Mit " wakeonlan -i 10.101.199.182 00:D8:61:FF:58:7E #N182-FJ-P558" geht das, aber nicht mit " linbo-remote"?! „Übernehmen“ meldet zwar „erfolgreich“, aber er startet nicht. Den Befehl auf den Konsole:
meldet „not a pxe host“.
… ja, das passiert manchmal.
linbo-remote nimmt ein anders tool zum aufwecken (ethertool statt
wakeonlan?).
Vorschlag zur Güte: man kann es auch so machen.
man gibt per linbo-remote einen Befehl der nur einmal beim nächsten
boot abgearbeitet wird.
Das ist
linbo-remote -p format,sync:1, … (geht auch mit -i und -g und -r)
(also statt linbo-remote -c nimmt man -p …).
Clients mit wakeonlan wecken: dann wir bei diesem booten die durch -p
hinterlegte commandfile abgearbeitet
Aber: es lohnt sich vielleicht auch mal die Switches stromlos zu stellen
für eine Minute und den server neu zu starten und dann nochmal
linbo-remote -w … versuchen.
Hm, also ehrlich gesagt, sollte das aus der Weboberfläche raus direkt funktionieren. Ich kann doch den Lehrern doch nicht auf solche Alternativen hinweisen. Wenn das dann funktioniert, was ich nicht getestet habe und auch vermutlich nicht werde. Es reicht mir schon, dass ich hier im Forum wühlen muss.
Hm, also ehrlich gesagt, sollte das aus der Weboberfläche raus direkt
funktionieren. Ich kann doch den Lehrern doch nicht auf solche
Alternativen hinweisen. Wenn das dann funktioniert, was ich nicht
getestet habe und auch vermutlich nicht werde. Es reicht mir schon, dass
ich hier im Forum wühlen muss.
… nun ja: WakeOnLan ist ein sehr komplexes Ding.
Dass das in manchen Settings Probleme macht ist bekannt.
Und das manchmal die eine Art die Rechner auf zu wecken besser ist als
die andere ist auch ein Erfahrungswert.
In meinen Einrichtungen hab ich z.B. wenig Probleme mit linbo-remote -w
Wobei in der Umgebung, die nur HP Switches und HP Clients hat das sehr
viel zuverlässiger funktioniert.
In der anderen mit gemischten Switches und Consumerhardware gibt es
eigentlich immer ca. 10% die halt nicht aufwachen (immer andere).
Ich habe noch nie erlebt, dass ein Lehrer administrative Rechte hat. Das war bislang nötig, um linbo-remote ausführen zu können. Hab ich da etwas verpasst?
Was ich noch festgestellt habe. Die angesprochene IP-Adresse ist nicht in dem Netz, welches wir üblicherweise verwenden. Vielleicht liegt es einfach daran, dass linbo-remote nicht funktioniert.
root@srv:~# /usr/sbin/linbo-remote -i N182-FJ-P558 -w 60 -n -p start:1
###
### linbo-remote (9259) start: Wed Jul 6 11:13:54 CEST 2022
###
Preparing onboot linbo tasks:
N182-FJ-P558 ... Done.
Trying to wake up:
N182-FJ-P558 ... Sending magic packet to 255.255.255.255:9 with 00:D8:61:FF:58:7E
Waiting 60 second(s) for client(s) to boot ............................................................
Verifying onboot tasks:
N182-FJ-P558 ... Not done, host skipped!
###
### linbo-remote (9259) end: Wed Jul 6 11:14:56 CEST 2022
###
erst recht nicht mit IP-Adresse:
root@srv:~# cat /etc/linuxmuster/sophomorix/default-school/devices.csv | grep N182-FJ-P558
N182;N182-FJ-P558;linux-efi;00:D8:61:FF:58:7E;10.101.199.182;;;;classroom-teachercomputer;;1;;;;;
root@srv:~# /usr/sbin/linbo-remote -i 10.101.199.182 -w 60 -n -p start:1
Skipping 10.101.199.182, not a pxe host!
Usage: linbo-remote <options>
Options:
-h Show this help.
-b <sec> Wait <sec> second(s) between sending wake-on-lan magic
packets to the particular hosts. Must be used in
conjunction with "-w".
-c <cmd1,cmd2,...> Comma separated list of linbo commands transfered
per ssh direct to the client(s).
-d Disables gui. To be used only together with option -c.
-g <group> All hosts of this hostgroup will be processed.
-i <i1,i2,...> Single ip or hostname or comma separated list of ips
or hostnames of clients to be processed.
-l List current linbo-remote screens.
-n Bypasses start.conf configured auto functions
(partition, format, initcache, start) on next boot.
To be used only together with options -p
or -c in conjunction with -w.
-r <room> All hosts of this room will be processed.
-p <cmd1,cmd2,...> Create an onboot command file executed automatically
once next time the client boots.
-w <sec> Send wake-on-lan magic packets to the client(s)
and wait <sec> seconds before executing the
commands given with "-c" or in case of "-p" after
the creation of the pxe boot files.
-u Use broadcast address with wol.
-s Select a school other than default-school
Important: * Options "-r", "-g" and "-i" exclude each other, "-c" and
"-p" as well.
Supported commands for -c or -p options are:
partition : Writes the partition table.
label : Labels all partitions defined in start.conf.
Note: Partitions have to be formatted.
format : Writes the partition table and formats all
partitions.
format:<#> : Writes the partition table and formats only
partition nr <#>.
initcache:<dltype> : Updates local cache. <dltype> is one of
rsync|multicast|torrent.
If dltype is not specified it is read from
start.conf.
sync:<#> : Syncs the operating system on position nr <#>.
start:<#> : Starts the operating system on pos. nr <#>.
create_image:<#>:<"msg"> : Creates a image image from operating system nr <#>.
upload_image:<#> : Uploads the image image from operating system nr <#>.
reboot : Reboots the client.
halt : Shuts the client down.
<"msg"> is an optional image comment.
The position numbers are related to the position in start.conf.
The commands were sent per ssh to the linbo_wrapper on the client and processed
in the order given on the commandline.
create_* and upload_* commands cannot be used with hostlists, -r and -g options.
No valid hosts in list!
mit
root@srv:~# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: ens18: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 02:46:9d:7b:72:d3 brd ff:ff:ff:ff:ff:ff
inet 10.0.0.1/16 brd 10.0.255.255 scope global ens18
valid_lft forever preferred_lft forever
inet6 fe80::46:9dff:fe7b:72d3/64 scope link
valid_lft forever preferred_lft forever
Wie gesagt:
root@srv:~# wakeonlan -i 10.101.199.182 00:D8:61:FF:58:7E #N182-FJ-P558
Sending magic packet to 10.101.199.182:9 with 00:D8:61:FF:58:7E
startet ohne Problem. Warum nicht „linbo-remote“???
server srv in VLAN 3 mit 10.0.0.1
client N182-FJ-P558 in VLAN101 mit 10.101.199.182
firewall opensense mit 10.0.0.254 (VLAN3) und 10.101.0.254 (VLAN101)
Das Netz der Lmn hat die Netzmaske 255.255.0.0. Das bedeutet m.W. Dass der Broadcast für wol auf der Adresse 10.0.255.255 raus geht. Die Rechner im falschen Netz werden deshalb nicht erreicht.
Ich kann doch die Clients nicht im selben Netz wie die Server laufen lassen. Dafür ist doch opensense da. Clients und Server im selben Netz wäre sicherheitstechnisch und auch schon wegen der Auslastung nicht gut. Daran kann es aber auch nicht liegen.
Ich frage nochmal:
Warum funktioniert wakeonlan?
Und linbo nicht?