Tech Work Notes

Installing Puppet on RHEL 5

2011-10-12T10:41:00.006-05:00

Install puppet on rhel5.

First install the EPEL repo from redhat.
http://fedoraproject.org/wiki/EPEL#Extra_Packages_for_Enterprise_Linux_.28EPEL.29

Then install puppet with yum.
yum install puppet puppet-server.

Quick way to keep servers similar with the same packages.

2011-05-26T15:57:00.005-05:00

Here is a quick and dirty way to keep RHEL servers updated with the same packages that are installed with yum.

Use the following command to create a text file with the packages installed.
rpm -qa > packagelist

Then you can copy the file packagelist to your new server. Run the following command to do a yum install of all the packages that are in the packagelist file

yum -y install $(cat /pathto/packagelist)

That should get and install all the same packages to the other server.

RHEL 5, NFS, and Firewall

2011-04-05T08:34:00.007-05:00

Setting up NFS on a RHEL 5 box is very simple and there are a few more configuration edits that you should make to allow the firewall to pass the NFS connections through.

First edit the /etc/sysconfig/nfs file.

Uncomment the following lines:
LOCKD_TCPPORT=32803
LOCKD_UDPPORT=32769
MOUNTD_PORT=892
STATD_PORT=662

Now restart NFS using RHEL you can use the service command
service nfs restart

Then you need to add those ports to your firewall. An easy way to do this is to edit the /etc/sysconfig/iptables file. If you are using RHEL then use the following lines. Otherwise edit the lines to match your firewall config.

-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 111 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 111 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 662 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 662 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 892 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 892 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 2049 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 2049 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 32803 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 32803 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 32769 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 32769 -j ACCEPT

Once these entries have been made restart the firewall. You can use the following command in RHEL to restart the firewall.
service iptables restart

Now test your NFS connection.

Windows printing with a CName

2010-11-04T11:16:00.004-05:00

Once you have the cname in the DNS server enter the following registry keys.

HKLM\system\currentcontrolset\services\lanmanserver\parameters
Add value
Value Name: OptionalNames
Data type: REG_SZ
Value Data: "alias.alias.com"

HKLM\System\CurrentControlSet\Services\LanmanServer\Parameters
On the Edit menu, click Add Value, and then add the following registry value:
Value name: DisableStrictNameChecking
Data type: REG_DWORD
Radix: Decimal
Value: 1

Then run the following commands

setspn -a host/aliasname targetserver

setspn -a host/aliasname.alias.com targetserver

Reboot the server.

Taken from the following technet article.
http://support.microsoft.com/kb/281308

RHEL 5 hald-addon-storage

2010-07-15T08:51:00.003-05:00

So my RHEL 5 VMs started to show high cpu utilization on a few machines. I noticed that the even though top wasn't showing high cpu utilization vmware was. I saw that the hald-addon-storage was polling /dev/hda pretty often. I determined that this was probably the cause of the high cpu alert. So to fix this issue I went to my VMs shut them down and removed the CD-ROM drive from the system. I then booted the systems back up and waited for them to finish. I then shut them down again and added the CD-ROM drive back to the VM. This seemed to resolve the problem with the hald-addon-storage polling too often and my VMware system is no longer alerting on the high cpu utilization.

EMC Engineering Mode

2010-05-04T17:17:00.001-05:00

To go into EMC Engineering mode you can press ctrl+shift+F12 then enter the password messner. Be careful what you do though once in this mode.

VMWare ESX 3.5 Host showing disconnected.

2010-04-26T15:55:00.003-05:00

I encountered this situation. I had a host machine that was showing disconnected in Virtual Center. The guest were also showing to be disconnected in Virtual Center. The guest were still up and running the host was also up and running. The host could be pinged but could not be logged into through the VIC. I ran a ssh session to the host. I then restarted the following service. This seems to have resolved the issue.

service mgmt-vmware restart

Vmware ESX Snapshots will not delete.

2010-04-14T15:30:00.008-05:00

If you have vmware snapshots that you do not remove frequently you may start to notice that the datastore that the guest is on will start to fill up. If you start to notice that these snapshots are eating disk space up one thing to do is delete the snapshots and consolidate the changes that you have made to the original vmdk file.

WARNING: Always have a backup just in case!

If you would like to delete the snapshots one thing is to make sure that you have enough disk space to do this. I recommend having about 2x times the amount of space the guest is using in free space to delete the snapshots. When a snapshot is deleted it takes the current changes and applies them to the original vmdk file.

Every time a new snapshot is made 2 files are created. One is the xxxxxx.vmdk and the other is xxxxx-delta.vmdk. The xxxxx.vmdk file is just a file that holds the information like disk geometry and other pointers in there that reference the delta file. The delta file is the actual disk where the data is stored.

If you deleted the snapshots and did not have enough free space then the following might have occurred. It halts at 95% complete and then fails or it completes the task but the vmdk files still exist. You will also notice in the VIC that the option to remove the snapshots is now gone. You can ssh to the host machine and navigate to the datastore that holds the guest. Usually this path is /vmfs/volumes// If you still have the xxxxx.vmdk and the xxxxxx-delta.vmdk file then you should be ok. We will verify that the guest is still pointing to the vmdk file with the most current information. Open a ssh session to the vm host. Navigate to the directory on the datastore where the guest lives. You can cat the .vmk file and you want to check for the following information.

scsi0:0.fileName = "-xxxxxx.vmdk"

This xxxxx.vmdk file should be the most current disk. If is not then you can shutdown the guest and change the file and point it to the newest snapshot file that you have. Some other things to check as well is to verify that the snapshot disks are all still linked together. To do this you can cat the .vmdk file and xxxxx.vmdk file and check the CID and parentCID. The parentCID should point to the original disk or the snapshot disk before. For instance: .vmdk will show the following:

CID=e6d85194
parentCID=ffffffff

The xxxxxx.vmdk file should show something like the following:

CID=41e19117
parentCID=e6d85194

If you have another snapshot then the next one will contain the following:

CID= e67d8ac1
parentCID=41e19117

As you can see the parentCID of the snapshot file should point back to the CID of the original disk. If this does not match then you should be able to start with the original disk then fill in the parentCID of each snapshot to point to the newer snapshot. Once this has been verified or changed then the system can be booted back up and should be ready to go.

This will not solve the the fact that you still have snapshots out there. But will verify that your virtual machine has not been corrupted by the process of trying to consolidate the disk by removing the snapshots.

One way to fix this is to use vmware converter on the virtual machine and do a hot clone of the guest. Once the machine is converted you will have only one vmdk file. You can then discard the original guest and use the new vm you have just created. Another way is to create a new virtual machine and do a file level backup and restore to the new virtual machine.

Flushing DNS Cache on Leopard 10.5 and Windows systems

2009-09-29T14:30:00.003-05:00

To flush the DNS Cache on windows systems use the following command.

ipconfig /flushdns

To flush the DNS Cache on a system running Leopard 10.5 use this command.

dscacheutil -flushcache

RHEL 5 Sendmail using to send mail to Relay MTA.

2009-09-29T13:00:00.001-05:00

So I have been working on sendmail so that we are not running the sendmail daemon in the background on RHEL 5 and just have the server submit mail to our main SMTP server. To do this edit the following files.

/etc/sysconfig/sendmail

change the line

DAEMON=yes
to
DAEMON=no

Then edit the following file.

/etc/mail/submit.cf

change the line

D{MTAHost}[127.0.0.1]
to
D{MTAHost}[smtp.relayserver.com]

Restart the sendmail service

service sendmail restart

That's it.

Print Spooler 100%

2009-09-28T10:50:00.000-05:00

Well I was working on a Windows print server today and found that the spooler service was pegged out at 100%. To resolve this stop the print spooler service.

GUI: start>run>services.msc Print Spooler right click stop
CMD: "net stop spooler"

Go to the following directory and delete all the files.

C:\WINDOWS\system32\spool\PRINTERS

Start the print spooler service back up.

GUI: start>run>services.msc Print Spooler right click start
CMD: "net start spooler"