Andrew Siwko’s Blog #2

I’m copying off Roller posts and preparing to decommission.  The date in a post will be the original posting date from Roller.

[linode ~]# uptime
00:04:42 up 2 min,  1 user,  load average: 0.02, 0.03, 0.00

My linode is undergoing maintenance.  I’ve got to migrate it to another box.

[linode ~]# uptime
23:38:54 up 713 days, 23:55,  2 users,  load average: 0.00, 0.00, 0.00

Pretty good!

I got another linode, copied the image to it and got it to boot by turning on the xen helper.

Now it’s time to migrate the original!

I’d like to have the Nagios watch over Bacula jobs and report if they have not been run successfully after a period of time.  I think I’m going to use a “run after” job but on the director, not on the client.  Then I can check the log age with a file age check plugin…

Done!

When an incremental backup is more than 18 hours old I get a warning.  When a full backup is more than 4 weeks old I get a warning.

I scripted the server cert creation and CA signing back in March of 2005. I needed a cert for this server and ran the script like this:

[root@belloc certs]# ./mk-server-cert.sh linode.siwko.org

One passphrase and one CN later, I’ve got certs for the new host! Automation saves time.

Background:

• Private network behind NAT gateway
  
• siwko.org domain with mix of public and private hosts
  
• public DNS hosted by zoneedit.com
  
• Existing DNS is bind on Fedora Core 2
  
• Local DHCP feeds DNS
  

Requirements:

• DHCP / DNS integration
  
• Management interface GUI/Web
  
• Management API (java preferred)
  
• Supports split public/private scheme
  
• Supports 2 private nodes for availability
  

Notes:

While researching, I stumbled on webmin which supports bind configuration.

Alternate products mentioned in posts:

• dnsmasq (http://en.wikipedia.org/wiki/Dnsmasq)
  
• djbdns (http://en.wikipedia.org/wiki/Djbdns)
  

Things are slow this morning:

09:32:06  up 46 days, 10:54,  1 user,  load average: 49.82, 46.76, 33.58

I don’t delete my old http server logs and analog is putting a huge load on the box processing them.

I’ve got some machines with CD-ROM drives (no DVD). I couldn’t find the CD image for Fedora 8, but I had one around for Fedora 7. So, after FC7 was installed, I googled about and found these two links:

• ioncannon.net
• fedoraproject.org/wiki

I had a bit of trouble on one of them, but I deleted gnupg2 (see this) and it worked fine after that.

I was able to install and upgrade to Fedora 8. Now I’m looking forward to Fedora 9.

 13:00:42  up 243 days, 23:21,  5 users,  load average: 74.59, 35.59, 13.68
240 processes: 239 sleeping, 1 running, 0 zombie, 0 stopped
CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
           total    2.4%    0.0%    3.1%   0.3%     1.2%    1.2%   91.4%
           cpu00    1.5%    0.0%    3.3%   1.5%     4.9%    0.7%   87.6%
           cpu01    3.1%    0.0%    2.5%   0.0%     0.0%    0.9%   93.2%
           cpu02    2.7%    0.0%    2.7%   0.0%     0.0%    1.3%   93.0%
           cpu03    2.3%    0.0%    3.9%   0.0%     0.0%    1.7%   91.8%
Mem:  2055464k av, 2036816k used,   18648k free,       0k shrd,   93068k buff
                   1567216k actv,  297144k in_d,   30172k in_c
Swap: 4194216k av,       0k used, 4194216k free                 1768936k cached

I passed the Red Hat Certified Engineer exam on Red Hat Enterprise Linux 4 on March 10th, 2006.