Bug 289730 – net-analyzer/nrpe /etc/init.d/nrpe reload makes daemon unreacheble

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Marcin Mirosław 2009-10-19 14:15:55 UTC]

After reloading of nrpe dameon, i can't connect to him anymore, but process is still alive.

Reproducible: Always

Steps to Reproduce:
1./etc/init.d/nrpe start
2.# /usr/lib/nagios/plugins/check_nrpe -H 127.00.0.1 -c check_hda1
DISK CRITICAL - /dev/hda1 is not accessible: Nie ma takiego pliku ani katalogu
3./etc/init.d/nrpe reload
nrpe            |* Reloading nrpe...                                                                                                  [ ok ]
marcinm init.d # /usr/lib/nagios/plugins/check_nrpe -H 127.00.0.1 -c check_hda1
CHECK_NRPE: Error - Could not complete SSL handshake.

[Comment 1 Marcin Mirosław 2009-10-19 14:59:44 UTC]

I found segfault in logs. After reloading nrpe, connecting to nrpe makes segfaulting in libwrap:
localhost nrpe[25161]: warning: cannot open /etc/hosts.allow: Permission denied
localhost kernel: nrpe[25161]: segfault at ffffffff ip b76fb252 sp bf8e4960 error 4 in libwrap.so.0.7.6[b76f8000+7000]

User nagios can't read file /etc/hosts.allow .
I'm not sure , should this bug be assigned to tcp-wrapper package?

[Comment 2 Diego Elio Pettenò (RETIRED) 2012-08-23 23:10:39 UTC]

I'm very tempted to just mask tcpd USE flag ... but more likely I guess we should disable the reload method when using it, as hosts.* are not readable..

[Comment 3 Diego Elio Pettenò (RETIRED) 2012-08-24 04:54:20 UTC]

Yep, removed reload in this situation.