[SOLVED] ruptime is not showing all the machines in network

someshpr · 06-03-2010, 04:30 PM

Hi,
I usually monitor machines in the network via ruptime. But one machine (lambda) acts very odd. It never shows up in the output of ruptime, unless ruptime is run from a particular client (delta).

For example, say I am logged in lambda. This is what the output of ruptime and few other commands are:

Code:

[root@lambda ~]# ruptime
alpha         up   9+04:18,     4 users,  load 0.01, 0.08, 0.08
beta          up   8+06:14,     3 users,  load 0.05, 0.23, 0.22
delta         up  41+20:52,     0 users,  load 0.16, 0.51, 0.31
eps           up   9+00:20,     2 users,  load 0.10, 0.04, 0.01
gamma         up      1:33,     1 user,   load 1.10, 1.03, 0.97
pi            up  27+01:07,     7 users,  load 1.55, 1.82, 1.54
sigma         up   1+03:56,     3 users,  load 0.01, 0.18, 1.71
tau           up 146+03:59,     1 user,   load 0.10, 0.24, 0.13

[root@lambda ~]# cd /var/spool/rwho/
[root@lambda rwho]# ll
total 32
-rw-r--r-- 1 root root 228 Jun  3 15:44 whod.alpha
-rw-r--r-- 1 root root 156 Jun  3 15:42 whod.beta
-rw-r--r-- 1 root root  84 Jun  3 15:43 whod.delta
-rw-r--r-- 1 root root 156 Jun  3 15:44 whod.eps
-rw-r--r-- 1 root root  84 Jun  3 15:42 whod.gamma
-rw-r--r-- 1 root root 252 Jun  3 15:42 whod.pi
-rw-r--r-- 1 root root 180 Jun  3 15:42 whod.sigma
-rw-r--r-- 1 root root 228 Jun  3 15:44 whod.tau
[root@lambda rwho]# uptime
 15:45:50 up 92 days, 22:36,  7 users,  load average: 0.02, 0.03, 0.00
[root@lambda ~]# uname -a
Linux lambda.grp.vpncc.org 2.6.18-164.11.1.el5 #1 SMP Wed Jan 6 13:26:04 EST 2010 x86_64 x86_64 x86_64 GNU/Linux

This is the case for any machine other than delta. None of them show lambda in ruptime result. But from delta it looks like this:

Code:

[root@delta ~]# ruptime
alpha         up   9+04:24,     4 users,  load 0.09, 0.05, 0.06
beta          up   8+06:23,     4 users,  load 0.44, 0.27, 0.22
delta         up  41+20:58,     0 users,  load 0.04, 0.25, 0.25
eps           up   9+00:26,     2 users,  load 0.08, 0.14, 0.07
gamma         up      1:39,     1 user,   load 1.12, 1.06, 1.00
lambda        up  92+22:40,     4 users,  load 0.09, 0.08, 0.02
pi            up  27+01:16,     9 users,  load 1.08, 1.18, 1.31
sigma         up   1+04:02,     3 users,  load 0.05, 0.08, 1.15
tau           up 146+04:08,     1 user,   load 0.13, 0.16, 0.12

[root@delta ~]# uname -a
Linux delta.grp.vpncc.org 2.6.18-53.1.21.el5 #1 SMP Tue May 20 09:35:07 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux

Only difference between delta and other machines is the fact that delta is the file server that hosts the nfs directories shared in all the machines. All the machines are in the same subnet. All of them run either RHEL4, RHEL5 or CentOS5. lambda runs RHEL 5.4, delta runs CentOS 5.2.

I tried restrting rwhod daemon on lambda, but no change.

Any idea why lambda is playing this hide and seek with ruptime?

If there is any other info I need to post, let me know.

TIA,

rweaver · 06-03-2010, 04:39 PM

One thing I'd like to mention at this point is, rwhod/ruptime/rwho (r-apps) are considered pretty insecure these days. You ~really~ shouldn't even be running these on machines behind a firewall imo. There are better solutions like snmpd and such.

However, addressing your problem directly-- I would say rwhod isn't running or accessible on lambda except on/to delta. May be a firewall issue or a hosts.* issue.

someshpr · 06-03-2010, 11:32 PM

Thanks, rweaver!

Quote:

Originally Posted by rweaver

One thing I'd like to mention at this point is, rwhod/ruptime/rwho (r-apps) are considered pretty insecure these days. You ~really~ shouldn't even be running these on machines behind a firewall imo. There are better solutions like snmpd and such.

May be I should switch to a more secure method, but first I need to figure out why is this issue with rwho in my current setup.

Quote:

Originally Posted by rweaver

However, addressing your problem directly-- I would say rwhod isn't running or accessible on lambda except on/to delta. May be a firewall issue or a hosts.* issue.

rwhod is running on lambda, as well as on all other machines. I even restarted it. But that didn't change the problem. The output is exactly same as lambda output if I run ruptime from alpha or any other machine except delta.

Can you elaborate on potential firewall/hosts* issue, please? Which places should I look for them?

TIA,

rweaver · 06-04-2010, 09:36 AM

Are you running iptables? Compare delta vs another machine, take a look at the config on lambda does it have anything special referencing delta? What do the hosts.allow and hosts.deny show?

someshpr · 06-16-2010, 08:56 PM

Sorry rweaver, I couldn't reply as I had to go out of country immediately after your reply, and will be out for another month. Added to that I am having trouble connecting to the cluster I had been writing about due to some vpn related issue from the new net connection I have. I think I'll be able to sort that issue soon, and then I'll get back with the status of the cluster iptables and hosts.* files.

TIA,

rweaver · 06-17-2010, 02:04 PM

No problem, enjoy your trip (even work trips out of the country can be enjoyable), if you still need help drop a line when you get back and we'll be glad to give you a hand.

someshpr · 11-18-2010, 10:18 AM

Quote:

Originally Posted by rweaver

Are you running iptables? Compare delta vs another machine, take a look at the config on lambda does it have anything special referencing delta? What do the hosts.allow and hosts.deny show?

I forgot to post this back! PROBLEM SOLVED!
iptables on delta was different than other machines. Once I cloned that iptable to other machines, everything is fixed.
iptable on each machine defines a group of machines which are allowed blindly. lambda was not part of that trusted group in any of the machines other than delta!

Thanks rweaver for the clue!