Nagios Update killed Nagios :(

[disrael]

So I decided to update Nagiosxi and it loads a page it just has these spinning/waiting logo displayed where it should be displaying content.

I noticed that there are i386 yum packages installed and for some reason apache was having issues restarting until I re-linked the apache modules to lib64. I don't know if these i386 packages are causing a conflict, nor not sure how they got on this system. Note we do have two required i386 packages:

HCLrootsh.i386 2.3-0 installed
glibc.i386 2.5-81.el5_8.7 installed

The above are installed on a system w/ Nagios XI that hasn't been upgraded and is working fine. We had these reinstalled after Nagios was installed.

These are the i386 packages installed on the Nagios system I tried upgrading:

apr.i386 1.2.7-11.el5_6.5 installed
apr-util.i386 1.2.7-11.el5_5.2 installed
audit-libs.i386 1.8-2.el5 installed
bc.i386 1.06-21 installed
cracklib.i386 2.8.9-3.3 installed
cyrus-sasl-lib.i386 2.1.22-7.el5_8.1 installed
db4.i386 4.3.29-10.el5_5.2 installed
device-mapper.i386 1.02.67-2.el5 installed
distcache.i386 1.4.5-14.1 installed
e2fsprogs-libs.i386 1.39-34.el5_8.1 installed
expat.i386 1.95.8-11.el5_8 installed
expat-devel.i386 1.95.8-11.el5_8 installed
glibc.i386 2.5-81.el5_8.4 installed
keyutils-libs.i386 1.2-1.el5 installed
krb5-libs.i386 1.6.1-70.el5 installed
libgcc.i386 4.1.2-52.el5_8.1 installed
libselinux.i386 1.33.4-5.7.el5 installed
libsepol.i386 1.15.2-3.el5 installed
libstdc++.i386 4.1.2-52.el5_8.1 installed
libtermcap.i386 2.0.8-46.1 installed
mailx.i386 8.1.1-44.2.2 installed
ncurses.i386 5.5-24.20060715 installed
openldap.i386 2.3.43-25.el5_8.1 installed
openldap-devel.i386 2.3.43-25.el5_8.1 installed
pam.i386 0.99.6.2-6.el5_5.2 installed
pcre.i386 6.6-6.el5_6.1 installed
perl-Net-DNS.i386 0.59-3.el5 installed
php-ldap.i386 5.1.6-39.el5_8 installed
postgresql-libs.i386 8.1.23-6.el5_8 installed
python-simplejson.i386 2.0.9-8.el5 installed
readline.i386 5.1-3.el5 installed
sqlite.i386 3.3.6-5 installed
sudo.i386 1.7.2p1-14.el5_8.4 installed
traceroute.i386 3:2.0.1-6.el5 installed
zlib.i386 1.2.3-4.el5 installed

[disrael]

Hrmm, this is very odd, so I looked closely in the apache logs and saw a javascript file getting a 403. I checked another server I have yet to upgrade and the permissions differed from the updated Nagios. For some reasons every file in: /usr/local/nagiosxi/html/includes/js was not readable to the world, which explains why it wasn't showing up.

Any idea why the update script would do this? Granted the update script error'ed out once or twice before I was able to finish it.

-Doron

[slansing]

What were the errors the upgrade script threw out when you tried the upgrade?

[disrael]

I think there was one with downloading an RPM and another one with the logging into the DB or something.

The odd thing is a lot of permissions are messed up. We installed the upgrade on two other boxes and those are fine as they were the previous revision. The revision we had installed on our test box was 2 or more revisions old.

Is there a script that I can run to reset all the permissions for Nagios. We keep finding ones that do not reflect the working boxes, so we are manually resetting them we we find it.

[slansing]

Yes, there is one located here though it may no longer work (appears to have for me):

http://assets.nagios.com/downloads/nagi ... ios_XI.pdf

After you fix the permissions run the following and recheck your web interface:

Code: Select all

service httpd restart

service nagios restart

[disrael]

It looks like it fixed some issues, but might have caused some others. Is it possible to just tar up everything on one of our other servers that the update went find on and untar it to this system? Should I tar up everything in /usr/local/nagiosxi/html and redeploy it to the new system?

-Doron

[mguthrie]

I would actually look at downloading the latest tarball, and running the install-html script. This will give you a fresh copy of all of the php files that get updated during an upgrade, and it sets permissions on the files as well.

[disrael]

Thanks, I tried it, though it didn't fix the issue we are facing. We get this error message when you click on Hostgroup Summary/Overview/Grid:

It appears as though you do not have permission to view information for any of the hosts you requested...
If you believe this is an error, check the HTTP server authentication requirements for accessing this CGI
and check the authorization options in your CGI configuration file.

[mguthrie]

Somehow I missed this earlier.

So I decided to update Nagiosxi and it loads a page it just has these spinning/waiting logo displayed where it should be displaying content

Right click in your browser and select "Reload", the old javascript files are being cached in the browser, I suspect it's preventing the page load.