poor man’s ruptime
Historically the original ruptime1 was using broadcast udp/5132 in a network. Since it's not 1982 anymore, but 2022 today, here's a version for multiple networks with encrypted traffic and client-server architecture.
You will automatically get instant list of hosts (down or up), inventory of hardware, software overview, comparable list of benchmark results.
While it was
rcp
(remote copy)rexec
(remote execution)rlogin
(remote login)rstat
ruptime
rwho
(remote who)rwall
(remote wall)
It is now
ruptime
(remote uptime) - the classicruname
(remote uname and OS/release) - keep track what OS/release you runrsw
(remote software) - what kind of package managers did sneak inrhw
(remote hardware, inventory) - what hardware you haverload
(remote load of CPU/MEM/GPU/GPUMEM) - usage of hardwarerbench
(remote benchmark) - comparable list of your hardwarernet
(remote network) - networking details (interface name, connection speed)rdisk
(remote disk) - overview of local disks and their speedsrac
(remote users' connect time) - overview of usage (see ac3)rwho
(remote who) - show who is logged in
The output shows how long the system has been up, the number of users currently on the system, and the load averages4. The load average numbers give the number of jobs in the run queue averaged over 1, 5 and 15 minutes.
$ ruptime # FQDN State Uptime Users Load Averages 1' 5' 15'
dolphin.ocean.net up 15+05:57 0 users load 0.04 0.08 0.07
fish.ocean.net up 4+21:27 0 users load 0.22 0.25 0.25
tuna.ocean.net up 4+21:27 0 users load 0.20 0.30 0.42
$ runame # FQDN Kernel Release Architecture, OS Version Code
banana.ocean.net Darwin 19.0.0 x86_64, MacOSX 10.15.1 19B77a
fish.ocean.net Linux 5.15.0-17-generic x86_64, Ubuntu 22.04 jammy
lemon.ocean.net GNU/kFreeBSD 11.4-0-amd64 x86_64, Debian unreleased sid
tuna.ocean.net Darwin 21.1.0 arm64, macOS 12.0.1 21A559
$ rload # FQDN CPU % MEM % GPU % MEM %
whale.ocean.net 19.00 3.37 51.20 42.12
$ rsw # FQDN pkg number...
seahorse.ocean.net dpkg 7243 rpm 0 pip3 393
$ rhw # FQDN age efi? cores memory
fish.ocean.net 2008/09/08 BIOS 8 31
lemon.ocean.net 2021/08/09 UEFI 16 16
banana.ocean.net 2019/12/10 UEFI 64 377
$ rnet # FQDN interface linkspeed wlanquality%
fish.ocean.net wlp3s0 144Mb/s 86
orca.ocean.net enp4s0 2500Mb/s
$ rbench # FQDN Memory Total CPU Cores
orca.ocean.net MEM 5.05 94 GB CPU 6.16 32
$ rac # FQDN hours users
orca.ocean.net 15122.94 4
$ rwho # user host date IP
root fish.ocean.net:pts/0 2023-06-12 07:28 (1.1.1.1)
me fish.ocean.net:pts/1 2022-06-23 15:54 (127.0.0.1)
you fish.ocean.net:pts/3 2023-06-28 06:33 (12.34.56.78)
we fish.ocean.net:pts/3 1995-06-28 06:33 (100.200.100.200)
us fish.ocean.net:pts/3 2023-06-28 06:33 (44.44.44.44)
-a Show all information about host
-h Help
-i Initialize the software
-m Multisite ruptime
-o Overview
-u Upload information to the server
-v Print license/version and quit
No option queries the server for the information.
- it's simple5
- monitoring systems have no or not very useful CLI tools
- you don't want to manually keep a list of hosts
- you want to see what hosts are down
- you want to see what hosts are not idle
- you want to run something on all running hosts with
parallel
- get rid of non-standard/in-house solutions that do not scale or are cumbersome in some other way
Get an overview of your operating systems and releases
$ runame | awk '{i[$NF]++} END {for (n in i) print i[n] " " n}' | sort -nr
Find hosts that are least used by CPU
$ rload | sort -k2n
Find hosts that have 90%+ usage of either CPU/MEM/GPU/GPUMEM
$ rload -c | grep " [9][0-9].\| [0-9][0-9][0-9]."
Update rnet
output for all online hosts
$ for a in `ruptime | grep -v " down " | awk '{print $1}'`; do echo $a; ssh root@$a "runame -u"; done
List all hosts sorted by network speed
$ rnet | sort -k3nr
Combined ruptime
and rload
output
$ join <(ruptime) <(rload) | column -t
Find missing reports
$ diff -y <(ruptime|awk '{print $1}') <(rhw|awk '{print $1}') | grep '<\|>\||'
Run something on all hosts having Ubuntu 22.04
$ runame | grep jammy | awk '{print $1}' | parallel -j0 'ssh root@{} "something"'
Get total cores and memory of all your machines
$ rhw|awk '{print $3 " " $4}'|datamash -t" " sum 1-2
Average age of computers, oldest and newest (by BIOS date)
$ rhw|awk '{print $2}'|sed "s,/.*,,g"|datamash -t" " median 1 min 1 max 1
Right adjusted rhw
output
$ rhw|column -t -R3,4
Your total diskspace
$ rdisk | sed "s,sd.,,g;s,nvme... ,,g;s,md.,,g;s,mmcblk.,,g" |sed "s,.*ethz.ch,,g" | awk '{for(i=t=0;i<NF;) t+=$++i; $0=t}1' |datamash sum 1
Number of users
$ rac | awk '{print $3}' |grep -v ^$ |datamash sum 1
483
Find hosts without ruptime
$ diff -y <(ruptime|awk '{print $1}') <(arp -a|sort|awk '{print $1}') | grep '>\||'
Sometimes nl
or ts
(from moreutils
) are useful as well.
To install the client run
apt install ruptime
To install the server run
apt install ruptimed
Installing on macOS
https://github.com/alexmyczko/homebrew-mac/blob/main/ruptime.rb
Installing on Alpine Linux
apk install cmd:lsb_release netcat-openbsd cmd:ac
The defaults for rwhod/ruptime is downtime after 11' (11*60 seconds)6 (ISDOWN), status messages are originally generated approximately every 3' (AL_INTERVAL)7.
SERVER=wedonthaveaprivacyproblem.com
PORT=51300
HOSTNAMECMD='hostname -f'
Create a key for the encryption with openssl
. You will need this on server and client for symmetric encryption.
COLUMNS=160 dd if=/dev/urandom bs=1 count=60 2>/dev/null > /etc/ruptime/ruptime.key
Create a local user to run the daemon.
adduser --disabled-password --quiet --system --home /var/spool/ruptime --gecos "ruptime daemon" --group ruptime
Running the daemon.
daemon --user=ruptime:ruptime mini-inetd 51300 /usr/sbin/ruptimed
If you set HOSTNAMECMD='hostname -s'
you will have the same mode as original rwho/ruptime/rwhod.
You can even limit the thing to your single one network with
iptables -A INPUT -p tcp --dport 51300 --match ttl --ttl-gt 1 -j REJECT
- Client:
nc
xz
bc
cron
memtester
timeout
wireless-tools
acct
- Server:
nc
xz
tcputils
daemon
- Optionals:
pen
trickle
bkt
iptables
wireless-tools
nvidia-smi
lm-sensors
- macOS
- Linux
- FreeBSD
-
FreeBSD: rc.d
-
Debian GNU/Linux: daemon, init.d, cron @reboot, systemd
-
macOS: https://launchd.info
-
without systemd
# crontab -l
*/1 * * * * /usr/bin/ruptime -u
*/3 * * * * /usr/bin/rload -u
@reboot /usr/bin/runame -u
@reboot /usr/bin/rsw -u
Some metrics are not useful to have at regular intervals, nor at every boot, so collect them when needed, examples:
rnet -u
On first setup and hardware changes (memory upgrade, disks added):
rbench -u
rdisk -u
rhw -u
Login to host and try these commands
ac -p
ethtool -i $(ifconfig|grep -v "^$"|sed "s,:.*,,"|grep -v "^ "|grep -v ^lo|head -1)
nvme smart-log /dev/nvme0n1
badblocks -sv /dev/sda
sensors
- https://github.com/rfmoz/tuptime
- https://github.com/dylanaraps/neofetch
- https://github.com/glpi-project/glpi-agent
- https://github.com/snipe/snipe-it
- https://github.com/PRTG/PythonMiniProbe/
- https://github.com/ganglia/monitor-core
- https://manpages.debian.org/unstable/manpages/nologin.5.en.html
- https://manpages.debian.org/unstable/manpages/issue.5.en.html
- https://manpages.debian.org/unstable/manpages/motd.5.en.html
- https://manpages.debian.org/unstable/proftpd-basic/ftpusers.5.en.html
- https://manpages.debian.org/unstable/cron/crontab.1.en.html references cron.allow cron.disallow
- https://manpages.debian.org/unstable/login/login.1.en.html
- https://manpages.debian.org/unstable/manpages/services.5.en.html
- https://manpages.debian.org/unstable/finger/finger.1.en.html