If you are ever tasked with updating the VMware tools in a linux guest from the command line, it is best to execute the installer from the console.  If you update the tools from a remote connection via ssh, the install will fail because the installer restarts the network service (unless you run the installer in the background by appending “&” to the end of the command).

After my mistake became clear, I opened vcenter and ran the installer from the console, hoping it would pick up where it stopped.  Some of the VMware tools files were still lingering in the system and installer displayed and error which stated it could not find vmware-uninstall.pl.

So here is the easy fix… simply remove the /etc/vmware-tools directory (or /etc/vmware) and re-run install-install.pl.  Just make sure to do it from the console. :)

Locked out of our own Blackboard system…

Posted April 10th, 2010 by admin

As you can imagine panic will set it when 10k users cannot access you system becuase they do not get a login screen.  We had an individual with Admin access inadvertently try to change the login screen without fully understanding the process. The result was a very useless error message, which made you believe it was a Tomcat issue.

If the page fails and you are unable to login in, enter the following address in your Web browser to bring up the default login page:

http://yoururl/webapps/login?action=default_login

You can then at least get to the Admin panel and reset the login to the default.

Every so often I will have problems when restarting Guest virtual machines on our ESXi servers.  One time in particular, I restarted the system from the operating system and the VM hung on reboot.  When I tried to hard Power Off the VM through ESXi, it gave me an error similar to “The operation cannot be performed in it’s current state”.  Here are some steps to take to regain control of the Virtual Machine… (none of these actions will affect the state of your running VMs)

Restart the Management services/agents; here is a great reference:

http://www.vm-help.com/esx/esx3i/ESXi_restart_agents.php

Run this command to tell inetd to re-read it’s configuration:

#  ps | grep inetd
#  kill -HUP “<pidofinetd>

Finally, follow the steps in this KB article:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1014165

This post is in reference to http://tracker.moodle.org/browse/MDL-10993

I had a little trouble finding the fix for this problem, and the resolution in the moodle bugtracker weren’t obvious.  So, to keep it short and sweet, here is the fix that I used.

The problem is a result of multiple restore processes running when the course import is executed.  This is caused by users who click the “(Continue)” link multiple times or they click it just before the page automatically redirects.  The link is displayed on the last page of the restore before it is finished.  Fix the problem by hiding the (Continue) link and replacing it with text to tell the user to wait.

add the following fw_layout.css for your current theme:

/* Hide Continue link */
#backup-restore #redirect #continue { display: none; }

add a message for the user by changing this line in moodle/backup/restorelib.php:

redirect($CFG->wwwroot.’/backup/restore.php?backup_unique_code=’.$backup_unique_code.’&launch=form&file=’.$file.’&id=’.$id, “Please wait…”, 1); }

If you are a Zimbra administrator and have an account setting that is incorrect, you might find that you just want to remove the setting completely or reset it to the default value.

You can list all the attributes for the account in question with this command:

zmprov ga <accountname>

Once you find the incorrect attribute, unset it like this:

zmprov ma <accountname> <attributename> ” (two apostrophes without a space between)

Droid or the Blackberry, decisons, decisions.

Posted March 20th, 2010 by admin

Our organization has been with Alltel for next to forever.  Old news Verizon gobbled them up, basically  swallowed them whole.  So in order to get over to the Verizon government plans (deeper savings, standard billing for line, etc). We need to replace our old phones (less than 3 months for some).  So I have been waffling back and forth, because most of our consultants and leadership team have been on Blackberry’s. They are reliable, sturdy, good battery life and sync well with our Zimbra Collaboration Server via the BES connector.

So I needed to make a decision, would the Blackberry grow with us as we move to a more digital organization?  I came to the conclusion it would not. The store for Blackberry apps seems empty in comparison to the Apple for iPhone or even the Marketplace for Droid.  Since the Droid is built on an open source linux flavor the apps will become more bountiful than the iPhone over time, I am certain.

These are the main reasons I opted for deployment of the Droid.

  • Uses ActiveSync to connect to Zimbra. Would no longer require the intermediate BES service.
  • Reduces the BES licensing Fee
  • Access to many new Applications that make our blogging and expanding our virtual presence easier. (more than the FB apps) Not as many as the iPhone but development will grow quickly.
  • It can multi-task unlike the Blackberry or iPhone. Meaning you can truly run two applications at the same time.  So if you are listening to a broadcast in one of our rooms and get an email you can check the email without losing the audio or video stream
  • it supports wi-fi to get on the Internet as well as the Verizon 3G network.

So soon the new phones will arrive and I will add an addendum to conversion and personal experiences. For know, I feel very confident these were the Droids we were looking for.

RDP Drive Forwarding, not visible

Posted March 18th, 2010 by admin

I have fought for several days trying to get drive forwarding to show on my Blackboard server. The client tool I use Cord for OSX is a lot like Remote Desktops in the Administration Kit. I have Drive Mapping on in the client and they never show on the server. I finally figured the problem need to be on the server.

So here is what you need to check.

Goto Start>>>>Administrative Tools>>>>Terminal Service Configuration


Right Click the RDP-TCP connection. View the properties.

Check the Disable Drive Mapping option. In my Case it was checked and that what was preventing my forwarding from working.

Hope this helps… Pete

FCKEDitor, Drupal and Firefox Spell Checker

Posted March 18th, 2010 by admin

Firefox has a great built in Spell Checker. However when setting up this site in Drupal, we chose to implement FCKeditor. The problem is FCKeditor uses its own menu system, so the right click context menu is not available for the Firefox spell Check.  Also the Firefox spell checker is not on by default with FCKeditor.   I have done some searches and found the following three resources.

To save you some reading the steps are very simple.

  1. FTP to your server
  2. /var/www/vhosts/{website}/httpdocs/sites/all/modules/wysiwyg/editors/js/fckeditor.config.js
  3. Add  two lines to the bottom of the file.
  • FCKConfig.FirefoxSpellChecker = true ;
    FCKConfig.BrowserContextMenuOnCtrl = false;
    
  1. Save the file and head over to edit a post in Drupal.

If all went well You will see red squiggles under your typos.

Also to get the context menu. You can hold the CTRL key and Left Click a link to see suggestions.

Sometimes I have to work on our linux servers (command line only) remotely using the console provided by the vcenter client for our ESXi servers.  I have to use vmware Fusion to run the vcenter client.  The problem is that you have to use the shift + page up/down buttons to scroll the console window up when there is too much text to fit in your viewable; and Mac laptop keyboards do not have these page buttons.  Here is handy shortcutf to alleviate this problem and some others that you might find useful.

fn + up/down arrows – functions as page up/down to scroll one page at a time and can be used in many applications

shift + fn + up/down arrows – scroll the console windows up and down in vcenter client

command + up/down arrows – navigate to the beginning or end of a document or web page

space/shift + space – functions as page up/down in document or web page

DRBD – Fix for Diskless status

Posted March 15th, 2010 by bryanr

Here is something I found very useful when working with DRBD.  After running some system updates on my Openfiler server that is running DRBD and Heartbeat, my logical volumes were showing up as “Diskless” when I looked at the DRBD status.  Looks something like this:

vg0             Connected       Secondary/Primary     Diskless/UpToDate    C

When I tried to run the attach command to sync up the volumes, I just received an error that said the disk could not be mounted.  And the error led me to believe the device was already mounted and in use.  I found that when the disk is mounted correctly through DRBD when it has the Primary role it will show up with a mount point here:

/dev/mapper/<volumename>

If the DRBD node is in the Secondary role, these folders will not be present.  If they are, it’s good evidence that something is wrong.  Run this command on the Diskless server and reboot.  Your volumes should then sync up like normal.

kpartx -d /dev/mapper/<volumename>