What is active?
One can find out what processes are normally running via ps ux:
...
| Code Block |
|---|
voyage:~# netstat -nlp Active Internet connections (only servers) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 0 0.0.0.0:111 0.0.0.0:* LISTEN 2830/portmap tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 3062/apache tcp 0 0 0.0.0.0:53 0.0.0.0:* LISTEN 6063/dnsmasq tcp 0 0 0.0.0.0:1723 0.0.0.0:* LISTEN 3006/pptpd tcp6 0 0 :::53 :::* LISTEN 6063/dnsmasq tcp6 0 0 :::22 :::* LISTEN 6032/sshd udp 0 0 0.0.0.0:53 0.0.0.0:* 6063/dnsmasq udp 0 0 0.0.0.0:67 0.0.0.0:* 6063/dnsmasq udp 0 0 0.0.0.0:111 0.0.0.0:* 2830/portmap udp6 0 0 :::53 :::* 6063/dnsmasq Active UNIX domain sockets (only servers) Proto RefCnt Flags Type State I-Node PID/Program name Path |
Finding things
The data is kept in /usr/local/share/pinger/data/ with files of the form ping-yyyy-mm.txt where yyyy is the year and mm the month (1-12).
The CGI scripts are kept in /usr/lib/cgi-bin:
| Code Block |
|---|
voyage:~# ls \-l /usr/lib/cgi-bin/local/share/pinger/data/ total 723404 -rwxr-xr-x-rw-r{-}{-}r-\- 1 root rootstaff 3480630 May 18 17:31 ping-2009-05.txt voyage:~# date Mon May 18 21:43:31 GMT 2009 |
The CGI scripts are kept in /usr/lib/cgi-bin:
| Code Block |
|---|
voyage:~# ls \-l /usr/lib/cgi-bin total 7215101 Jan 7 2000 cgi-lib.pl -rwxr-xr-x 1 root root 15101 Jan 7 2000 cgi-lib.pl -rwxr-xr-x 1 root root 20101 May 15 23:17 ping_data.pl -rwxr-xr-x 1 root root 36226 May 12 22:00 traceroute.pl |
Updating crontab
Edit /etc/crontab and save. then reboot to make the new version take effect (alternatively use kill -HUP <PID> (e.g. kill -HUP 3039) to force an update of the crontab process, where you can find the PID from the ps ux command to restart the crontab process- see above).
Logs
There is an example of a crontab:
Logs
The log is The log is kept in /var/log/messages. We believe this is "rotated" (compressed, saved and deleted) automatically by syslog using /etc/logrotate.conf and /etc/syslog.conf, so memory is not eventually exhausted.
Applying Patches
On the one hand one wants to keep the system and applications patched to the latest releases. On the other hand this may require considerable bandwidth to download the patches, or timeouts, it could gobble up memory in the host etc.
To find out how much memory is available in the file system, use df -k:
| Code Block |
|---|
voyage:~# df \-k Filesystem 1K-blocks Used Available Use% Mounted on rootfs 3853536 249880 3407904 7% / udev 10240 20 10220 1% /dev /dev/disk/by-label/ROOT_FS 3853536 249880 3407904 7% / /dev/disk/by-label/ROOT_FS 3853536 249880 3407904 7% /dev/.static/dev tmpfs 128416 0 128416 0% /lib/init/rw tmpfs 128416 0 128416 0% /dev/shm tmpfs 8192 412 7780 6% /rw |
...
| Code Block |
|---|
voyage:~# apt-get update Get:1 [http://security.debian.org] etch/updates Release.gpg [197B] Hit [http://security.debian.org] etch/updates Release Ign [http://security.debian.org] etch/updates/main Packages/DiffIndex Hit [http://security.debian.org] etch/updates/main Packages Get:2 [http://www.voyage.hk] ./ Release.gpg [189B] Hit [http://www.voyage.hk] ./ Release Ign [http://www.voyage.hk] ./ Packages/DiffIndex Ign [http://www.voyage.hk] ./ Packages Hit [http://www.voyage.hk] ./ Packages Get:3 [http://ftp.hk.debian.org] etch Release.gpg [394B] Hit [http://ftp.hk.debian.org] etch Release Ign [http://ftp.hk.debian.org] etch/main Packages/DiffIndex Hit [http://ftp.hk.debian.org] etch/main Packages Fetched 3B in 6s (0B/s) Reading package lists... Done |
One applies the patches using apt-get update. However this can take over 45KBytes of memory. We have not decided if or how to automate the patch process.
Testing
Look at the last date the current file in /usr/local/share/pinger/data/ was updated and compare this to date, the times should be within 30 minutes:
| Code Block |
|---|
voyage:~# ls -l /usr/local/share/pinger/data/
total 3404
-rw-r--r-- 1 root staff 3480630 May 18 17:31 ping-2009-05.txt
voyage:~# date
Tue May 19 01:15:38 GMT 2009
|
If not see if pinger2 is executing properly:
| Code Block |
|---|
voyage:~# cd /usr/local/share/pinger
getIPV4: voyage
BeaconList is still up to date.
Going to sleep for 29.035192998008 minutes.
pingAllHosts(): Starting to ping all hosts.
pingHost(): Enter.
queryDNSCache: www.univ-batna.dz
getIPV4: www.univ-batna.dz
pingHost(): www.univ-batna.dz.
doPing(): Expanded pingCmd resolved to '/bin/ping -n -w 21 -c 1 -i 1 -s 56 193.194.69.34'
pingHost(): Enter.
queryDNSCache: www.ucan.edu
getIPV4: www.ucan.edu
pingHost(): www.ucan.edu.
...
|
Look at the date stamps on pingerCronStat.* and look at the contents if there is anything there:
| Code Block |
|---|
voyage:~# ls -l /usr/local/share/pinger/pingerCronStat.*
-rw-r--r-- 1 root staff 0 May 18 10:30 /usr/local/share/pinger/pingerCronStat.stderr
-rw-r--r-- 1 root staff 52934 May 18 11:02 /usr/local/share/pinger/pingerCronStat.stdout
|
Check the cronjob is running
| Code Block |
|---|
voyage:~# ps ux | grep cron
root 3039 0.0 0.3 2192 772 ? Ss May18 0:00 /usr/sbin/cron
root 4283 0.0 0.1 1640 508 pts/0 R+ 01:43 0:00 grep cron
|
If pinger2 is running (e.g. soon after the 30 mins interval in which it is scheduled) you might see something like:
| Code Block |
|---|
voyage:~# ps ux | grep pinger2
root 5613 0.0 0.4 2400 1060 ? Ss 03:00 0:00 /bin/sh -c cd /usr/local/share/pinger; \
perl /usr/local/share/pinger/pinger2.pl > /usr/local/share/pinger/pingerCronStat.stdout 2> \
/usr/local/share/pinger/pingerCronStat.stderr
root 5614 0.2 2.5 9048 6604 ? S 03:00 0:02 perl /usr/local/share/pinger/pinger2.pl
root 6091 0.0 2.1 9048 5408 ? S 03:16 0:00 perl /usr/local/share/pinger/pinger2.pl
root 6105 0.0 2.1 9048 5408 ? S 03:16 0:00 perl /usr/local/share/pinger/pinger2.pl
root 6157 0.0 0.1 1640 512 pts/1 R+ 03:17 0:00 grep pinger2
|
We verified that the host recovered after a power cycle.
- It pinged so the host was up,
- http://140.105.28.27/cgi-bin/ping_data.pl responds so Apache works.
- Submitting the request fetched appropriate data (i.e the latest was current within 30 mins, note the time-stamps are in Unix epoch time and are GMT.
Configuration
Logging onto the Host
Give yourself an IP in the same network as the ePinger machines. For example, you can give yourself 140.105.28.30 with netmask 255.255.255.128. You can then connect with any network cable to the Alix board and follow the instructions you already have. Of course, keep in mind that if you change the IP of the eth port on the right (as seen from the back) then you need to be on the same network to communicate with it.
Ethernet
Here is how to setup the network interface on the ALIX board:
- login as root
- remountrw
- cd /etc/network/
- vi interfaces
- just copy what is already there for eth0 and change the IPs.
- /etc/init.d/networking restart (this will restart the networking
services) - ifconfig (to check if everything is OK).
That should be it.
Keep in mind that eth0 is the plug on the right (there is sticker too), while eth1 is the one on the left.
After reboot, it takes some time before you are able to ssh. Don't panic (I did).
When you setup eth1, you should disable eth0. In this way it answers to ssh much faster. And at then end, they don't use both eth ports. On the other side, it's a good idea to have both for local troubleshooting.
Apache Logs
The Apache log rotation should be set up. This section just explains it a bit more:
1) check apache logs files:
ls /etc/log/apache/*.log
2) change directory:
cd /etc/logrotate.d
3) edit file: "apache" and change first line: /etc/log/apache/*.log
(default is /var/log/apache/*.log)
4) run: logrotate -f /etc/logrotate.conf
5) check that logs have rotated:
ls /etc/log/apache/.