Got Burggled
by OKButton on Jun.23, 2009, under Blogroll
In the last few weeks I have had my house broken in and I have been robbed. I have had my Mac and my camera stolen as well as my car. I am slowly getting all the insurance sorted out so I should be back up to speed again soon. The site has been broken for ages this seemed to be a nigally issue with fastcgi. This has been fixed with an update to debian.
Sharing Vialations within SYSVOL
by OKButton on Feb.24, 2009, under AD, Computing, Windows
I recently had an issue with sharing violations benig reported by Ultra Sound with issues within my sysvol replica set. I had a look at the tech net article on this problem (kb822300) but it did not help with the problem I was having as when I installed the surgested hotfix the files that were in a locked state were no loged in the event log as advertised. I decided to use the old fashond way to fix this.
First I ran:
ntfrsutl inlog > c:\sometextfile.txt
This creates a list of the items that are being inbound replicated on the problem server. I went through the file looking for the line IBCO_INSTALL_RETRY as described in the KB but it did not find it. Then I decided to go through the file manually reading the hole thing. I found that I had a slight variation on the KB article on some items I had the files marked with IBCO_DEL_RETRY and IBCO_RENAME_RETRY. I then ran the GUID’s for these files through the following command to find out what they were.
ntdfsutl idtable {GUID} > c:\someothertextfile.txt
This relieved that some folders that had been deleted would not replicate off the server and also a file that had been renamed would not replicate the change. I found that I had a permission issue with these folders and files and was able to fix it.
My Single Line WordPress Upgrade
by OKButton on Feb.23, 2009, under Computing, Hosting, HowTo, Linux, Web
Here is my take on the single line word press upgrade I saw at hack a day.
The bellow will download the latest version and unpack it. IT then removes the new plugins and themes as to not over wright your installed ones. IT then copies the new install over the old one and resets the permissions on the files.
wget http://wordpress.org/latest.zip && unzip latest.zip && rm -rf ./wordpress/wp-content/ && cp -r ./wordpress/* ../
You may need to change the paths for your installation but this should work quite well.
As allways take a backup before you run this just in case.
Designing an AD sites and services structure in an MPLS network
by OKButton on Feb.23, 2009, under Computing, Windows
So I have been looking into designing an AD replication structure for a network which is designed around all sites connecting into a managed MPLS network. So really all sites just connect to the MPLS cloud and data comes out the other end the routing of the cloud was not under my control. All the sources I have found on the net talk about sites being linked to other sites via leased lines and to map your AD design to the network design. This doesn’t really work in an MPLS network as all sites are connected to each other. So in this situation I believe that link cost can be treated a little differently.
First I decided to have two types of links core links with a cost of 10 and default site links with a cost of 100. Now I decided to put all my core sites (HQ’s and sites with large user populations) in to one site link called the hub. It should be noted that the KCC will fully mesh all sites that are in one link so if you find you are going to have alot of core hub sites you may want to construct the mesh manually by creating individual site links for the core. I would say more that 5 core sites probably requires manual link creation.
Next you need to create site links for all of your other sites. It is recommended to only have 2 sites in a link (apart from the core) this means that you can have full control of your replication paths. If you require backup links you can add these using a link cost of say 200. I decided to have sites that connected to the same MPLS core networks link together. This means that your replication traffic will only traverse the network via the core hub sites and not all the other sites. This cuts down on unnecessary bandwidth usage.
You should have no more than default satellite sites connected to one hub site. If you do you ether need to make sure the hub site has the resources in bandwidth and server performance to take the load or create another hub site and move some sites across to the new site.
You will also need to decide on replication intervals and schedules. This can be decided by how often you make changes and how quickly you need them to replicate.
To monitor your replication system I would recomend installing Ultrasound from Microsoft it is free and will monitor and debug any issues you might get in the future.
MacRumours Hacked
by OKButton on Jan.06, 2009, under Computing, Security, Web
The live feed for the MacWorld keynote from MacRumours was hacked today. This hack also took down there regular site. From the text that was in replacement to the keynote info (provided by thoes that had hacked the site) it seems that access to the server controle pannel was hacked allowing access to the hole system. It also apeared that more than one hacker had had ago at the system as conversations emerged in the streem. At one point it seemed that other users of the site were also able to contribute as requests to stop were also included in the streem as well as info from other streem providers. Not a good day for MacRumours.
