System uptime without ssh login

Just create a "xinetd" service, like that:

# cat /etc/xinetd.d/uptime service uptime { port = 10200 disable = no socket_type = stream protocol = tcp log_on_success += USERID log_on_failure += USERID server = /usr/bin/uptime user = root instances = UNLIMITED wait = no log_type = SYSLOG daemon debug } # service xinetd restart # netstat -ant | grep 10200 tcp6 0 0 :::10200 :::* LISTEN # telnet localhost 10200 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. 08:59:39 up 23 days, 16:36, 1 user, load average: 0.00, 0.01, 0.05 

If you have many servers, and I assume you will not just be interested in uptime, then the correct tool would be to load SNMP on each and have a monitoring station set up for their snmptrap responses. This is what the major players use. All the methods shown do not show missed boots, but SNMP will.

I would not fulfill your precept "without ssh login", but maybe it is a possible botched solution.

My idea is to make a script to which you pass the list of IPs that you want to check, so that you automate the process of logging in to each machine and asking for the uptime.

Something like:

#!/bin/bash if [ -z "$1" ]; then echo -e "" else read -a words -d "" < $1 for ip in "${words[@]}" do if ping -c1 $ip &> /dev/null; then ssh $ip -t "echo \$(hostname): \$(uptime)" 2>/dev/null; else echo "No communication with $ip" fi done fi 

Where the file it receives as an argument would be:

192.168.1.1 192.168.2.1 192.168.3.1 192.168.4.1 192.168.5.1 192.168.6.1 192.168.7.1 192.168.8.1 192.168.9.1 192.168.10.1 

The result would be like:

$ ./uptime-script.sh lan pc01-pc: 09:13:33 up 34 days, 1:40, 1 user, load average: 0.10, 0.16, 0.13 pc02-pc: 09:13:35 up 42 days, 2:50, 1 user, load average: 0.09, 0.10, 0.09 pc03-pc: 09:13:38 up 33 days, 23:37, 1 user, load average: 0.32, 0.19, 0.12 pc04-pc: 09:13:42 up 42 days, 2:15, 1 user, load average: 0.09, 0.13, 0.09 pc05-pc: 09:13:45 up 42 days, 2:08, 1 user, load average: 0.32, 0.25, 0.16 No communication with 192.168.6.1 pc07-pc: 09:13:58 up 42 days, 2:28, 1 user, load average: 0.20, 0.20, 0.18 pc08-pc: 09:14:01 up 6 days, 15:11, 1 user, load average: 0.77, 0.26, 0.14 server01-pc: 09:14:03 up 30 days, 3:02, 1 user, load average: 0.02, 0.05, 0.00 server02-pc: 09:14:05 up 30 days, 3:00, 1 user, load average: 0.01, 0.02, 0.00 

It would be a simple way to monitor these computers.

Very late to the party, but (assuming that uptime is not the only thing of interest in your vast server fleet's health) I'd suggest using a proper tool for that job.

Consider using Prometheus, scraping all sorts of metrics from your fleet in 5-minute intervals, including metrics like uptime.

Well. To query a machine's uptime you'd go node_time() - node_boot_time() (that's uptime in seconds, feel free to do the maths for different time units), but that's essentially what you're asking.

Also try rup. In debian packages rstatd on the server and rstat-client on the client.