AIX

I've set up errpt to redirect to syslog and syslog to forward to our linux syslog server. So I get to play around the rest of the day breaking my labs.

I've been filling file systems, tearing apart vhosts and unplugging cables - committing all sorts of heinous crimes.

I'm not used to this much freedom, I need some inspiration. What's your favorite break/fix?

bugs? did you not know - AIX is bug free.

just type $ bugs

Hard to track issue: /etc/hosts lacks entry for localhost and/or the /etc/hosts entry for the hostname has the wrong IP address.

Another syslog "source" is to activate ipsec4 and IP logging - runs on local4.*

And look at auth.* for login issues, smtp "logins" (I think they go there).

mail.* has many reporters (e.g., smtp for sure, pop3, etc.) and daemon.*

just turning them on separately and learning how they report (differently) can be interesting.

Hope this helps!

Login issues have been a good one.

Something I'm noticing, and maybe it was set up this way where I am before my time, if I fill a filesystem a la:

Code:

dd if=/dev/zero of=./blah bs=1M

I'll get my error in errpt. If I delete the file and do it again right away I don't get a second error. There seems to be a 'cool down' but I've seen instances in the past with flapping adapters where errpt rolls because it gets flooded.

That was the main reason for hunting down new ways to break the system without outright killing it.

Moderator's Comments:

Use code tags please.

Do not know about "cooldown" - and would need to know which daemon, subsystem is sending message to syslog and/or errpt before even beginning to guess.

My crude test would go something like this:

fill fs
see alert hit errpt
delete file I used to fill fs
fill fs again
no error hits errpt

Code:

#cd to the unlucky fs
cd /opt

#fill fs
dd if=/dev/zero of=./blah bs=1M

#verify error hit errpt
errpt -a | more

#del file
rm blah

#fill fs again
dd if=/dev/zero of=./blah bs=1M

#verify error hit errpt
errpt -a | more

Only one JFS2_FULL message in errpt for /opt, no entry for the second filling.


I redirect errpt to syslog using an additional to the ODM with this bad boy:

Code:

errnotify:
en_pid = 0
en_name = "errptToSyslog"
en_persistenceflg = 1
en_method = "/usr/bin/logger Error Log: `/usr/bin/errpt -l $1 | /usr/bin/grep -v TIMESTAMP` "

and add the appropriate syslog.conf entries to ensure I have the forwarded logs going the files/server I want and get them rolling (touch the files,

Code:

stopsrc -s syslogd; startsrc -a -n -s syslogd

).

I think the process goes:
*Error Hits* -> errpt entry -> errpt sends to syslog

Sorry if I'm throwing info out there, I figured I'd share what I'm doing in its completeness.