Dec 06

Windows 2008 DHCP Error 1046

I was trying to install and configure DHCP services on a sub-domain in my AD environment. After installing the role for DHCP I attempted to authorize the service. At first the snap-in stated that the authorization was successful. But under further investigation the IPv4 and IPv6 scopes were not being activated. In fact, the DHCP service itself was starting and then immediately stopping. When I attempted to re-authorize the service I received an “access denied” message. I saw some funky errors in the logs regarding how the system could not find the Enterprise server. I then logged in as a root domain Enterprise Admin, reinstalled the DHCP role, and attempted to authorize the service, this did not work either. But now I was at least getting a new error in the logs, a 1046

The DHCP/BINL service on the local computer, belonging to the Windows Administrative domain XXX, has determined that it is not authorized to start. It has stopped servicing clients. The following are some possible reasons for this: This machine is part of a directory service enterprise and is not authorized in the same domain. (See help on the DHCP Service Management Tool for additional information).
This computer cannot reach its directory service enterprise and it has encountered another DHCP service on the network belonging to a directory service enterprise on which the local computer is not authorized.
Some unexpected network error occurred.

At quick google of the issue turned up a MS post . The instructions stated to use the Manage Authorized Servers dialog to actually perform the authorization instead of just using the context menu. In the dialog I deleted the entry for the local DHCP server and added it back in to authorize. For some reason this did the trick and the DHCP service immediately started. Weird.

Aug 25

VMware vSphere 5 Now Available

VMware has released for download vShpere5, vCenter5, vShield5, VDR2, and VSA1. Sadly no vCD 1.5 yet :( . See you local VMware download section to get it.

Jul 19

Blogging VMWorld 2011

I plan on blogging my sessions from VMWorld 2011 this year. I am mainly taking advanced vCD, network, and performance/troubleshooting sessions. If I have some time I may try to fit in a lab or two, they were actually really good at PEX this year.

Jul 19

There is already another simpana running

When trying to start Commvault Simpana 8 on Linux (Redhat/CentOS) gives the error:
*** There is already another simpana running

Seems that if the CV instance dies it may not clean up after itself. Delete the lock file it keeps to allow you start the instances again.
[root@serv /]#rm -rf /tmp.lock_Galaxy
[root@serv /]#/etc/rc3.d/S99Galaxy start
[root@serv /]# ps ax | grep -i cvd
18309 ? SNsl 0:00 /opt/simpana/Base/cvd

Jun 15

VMware ESXi and Cisco EtherChannel

Today I was tasked with replacing one of the vSphere4 servers in my production cluster. Easy enough, get a fresh copy of ESXi installed on a new Dell R710. Reconfigure all the network settings to match the one it is replacing. Remove the current one from the cluster then shut it down and remove it from vCenter. Rack the replacement server and boot it up, and….. nothing….

I have no network connectivity, aargh. So I go down the list, check the cables, check the switch settings, check the vSwitch settings. Yup, port-channel mode on, yup Route based on IP hash. But still, only one of the two links on the management network work at any given time, plug the second one in and nothing.

At this point if I had hair I would be pulling it out. To the Google machine! And looky here, VMware has a KB Article on it. So using vCenter to set the settings on each portgroup did not work either. I enabled SSH for ESXi, connected, and ran the command line versions. It worked. I then ran it on my other vSwitches and portgroups.

Although the KB does not say why, I did rename my Service Console portgroup prior to changing the LB method. So I guess there is a glitch in the GUI that prevents the settings from filtering down correctly from the vSwitch.

May 05

Estimate System Availability

Taken in part from Cloud Application Architectures by George Reese.

The mathematical formulation of the availability of a component is:
a = (p -(cxd))/p
where:
a = expected availability
c = the % of likelihood that you will encounter a server loss in a given period
d = expected downtime from the loss of the server
p = the measurement period

The availability of a system is the total time of a period minus the sum of all the expected downtime during that period, all divided by the total time in the period:
a = (p - SUM(c1xd1:cnxdn))/p

Apr 29

Site Update

Just added a reading library section. Hopefully I can find the time to write reviews on them.

Apr 28

Patching ColdFusion 9

Just wanted to say… thanks Adobe for not putting any money into this platform and for hiring lazy product managers. It should not be this difficult to a)Find out what needs to be patched and b)Figure out how to actually patch a product. Really?? The 9.0.1 updater has been out for how long? And yet it is still not rolled into the base installation! Manual file modifications are needed in order to patch!

Rant over……

Okay, patching ColdFusion is a science and not for the faint of heart. There are so many things going on it will make your head spin. CF9 is not that bad, yet. A later post will cover CF8 for the non upgraders out there and it will be much more difficult. So, lets get started. This will cover the exact steps to get an installation up to the most recent hotfix and security hotfix level.

  1. Install ColdFusion 9.0 (if not already)
  2. Install ColdFusion 9.0.1 Updater
  3. Install CF9.0.1 Cumulative Hot Fix 1
    1. Make sure you follow the directions as there is more to patching than just the jar file.
    2. The CHF1 includes all hotfixes up to 2-22-2011
  4. Install Security Hotfix APSB11-04 (XSS fix)
  5. Update the JVM to the newest available
    1. Get the newest 1.6  JDK for your platform from Sun
    2. Install the JDK
    3. Edit C:Coldfusion9runtimebinjvm.config
      1. Near the top, comment out the line: java.home=C:/ColdFusion9/runtime/jre
      2. Add a new line under it: java.home=C:/Program Files/Java/jdk1.6.0_XX/jre
    4. Restart CF

That’s it, kinda painless unless you have a bunch of servers and/or instances to manage. In that case a simple script could automate the installation of just about everything.

Apr 28

Cacti Died

Noticed today that my Cacti installation was no longer updating it’s graphs. Ran the poller by hand from the command line:

[root@mon]# php /var/www/html/cacti/poller.php
[root@mon]# File size limit exceeded

Interesting… seems that there is a 2GB limit of the cacti.log file and when it hits that limit the poller refuses to run. So, delete/rename/move /var/www/html/cacti/log/cacti.log. The poller will start working immediately. FYI, i am running version 0.8.7d

Aug 26

Apache and Ubuntu Server 9.10

https://help.ubuntu.com/9.10/serverguide/C/httpd.html

Older posts «