Comments on: Automating vSphere ESX4 host installations

By: Land of the Lemmings, v2 » VMware vSphere 4 Environments

Land of the Lemmings, v2 » VMware vSphere 4 Environments — Tue, 16 Mar 2010 03:27:15 +0000

[...] Upgrading the existing ESX 3.5 nodes to ESX 4 is on the schedule for later this semester or early in the summer. To make the reinstall easier, I’ve scripted out the installs just about as much as possible using some excellent examples: Cylindric.net and Ubiquitous Talk. [...]

By: Manoj

Manoj — Mon, 08 Mar 2010 05:09:28 +0000

Hello Mike,

I’d like to thank you very much for pointing out where I was going wrong. You are a really an expert at bash scripting. I have always been very envious of people who have this ability. I have always wondered how one could learn to do this?

Keep up the great work.

By: Mike La Spina

Mike La Spina — Fri, 05 Mar 2010 12:46:52 +0000

Manoj,

The vimsh entry you are using appears to be a read only informational function. This configuration element is part of the /etc/vmware/esx.conf startup parameters.
You should have a line that resembles the following:

/boot/memSize = “640″

This value determines what the hypervisior will reserve for an esx console VM. It is this configuration element that must be set.
Your post script should find and replace the = memory value
You could use the following:

/bin/sed -i ’s/memSize = \”272\”/memSize = \”800\”/g’ /etc/vmware/esx.conf
/usr/sbin/esxcfg-boot -g
/usr/sbin/esxcfg-boot -b

Regards,

Mike

By: Manoj

Manoj — Thu, 04 Mar 2010 20:33:38 +0000

Hi Mike,

Sorry to trouble you, I have tried several things to try and get the console memory setting to 800MB but for some reason it still doesn’t make the change permanent even after 3-4 reboots. The message still indicates on next reboot.

Any pointers will be greatly appreciated.

Thanks
Manoj

By: Mike La Spina

Mike La Spina — Thu, 04 Mar 2010 18:07:19 +0000

Well done Kevin and thanks for sharing your experiance as well.
Regards,
Mike

By: Kevin Brooksby

Kevin Brooksby — Thu, 04 Mar 2010 10:39:39 +0000

Just thought I would provide an update and a sincere thanks for the hard work and time you put into documenting and sharing this with the community. I was able to resolve my issue with the my %post woes after initial kickstart reboot setting not sticking by adding a ‘/bin/sleep 120′ statement right after the #!/bin/bash line in the second %post section of the script. Everything is working perfectly and I am a dangerous man now ready to step my Virtual Infrastructure automation game up

By: Kevin Brooksby

Kevin Brooksby — Wed, 03 Mar 2010 03:28:40 +0000

I did not get a chance to try that but did come to the understanding that I needed another post section which I was missing originally so the /usr/bin/vmware-vim-cmd hostsvc/vmotion/vnic_set vmk0
and /usr/bin/vmware-vim-cmd hostsvc/memoryinfo 838860800 statements were not applying and throwing that rui.key ssl crypto error.

Here is my current post section, everything applies now (vSwitch related and all so i am further now) except that last post EOF section. Thoughts/tips would be greatly appreciated.

%post –interpreter=bash

/bin/sleep 90
/usr/sbin/esxcfg-auth –enablead –addomain=domain.local –addc=domain.local

/usr/sbin/groupadd -g 5000 lg-esxsu
/usr/sbin/useradd -u 501 -G lg-esxsu super1
/usr/sbin/useradd -u 502 -G lg-esxsu super2
/usr/sbin/useradd -u 503 -G lg-esxsu super3
esxcfg-firewall -e ntpClient
echo restrict default kod nomodify notrap noquerynopeer > /etc/ntp.conf
echo restrict 127.0.0.1 >> /etc/ntp.conf
echo server 10.20.0.200 >> /etc/ntp.conf
echo driftfile /var/lib/ntp/drift >> /etc/ntp.conf
/sbin/chkconfig –level 345 ntpd on
/etc/init.d/ntpd start
/usr/sbin/esxcfg-vswitch -A LAB_VMs vSwitch0
/usr/sbin/esxcfg-vswitch -A VMotion vSwitch0
/usr/sbin/esxcfg-vswitch -p VMotion -v 10 vSwitch0
/usr/sbin/esxcfg-vswitch -p LAB_VMs -v 11 vSwitch0
/usr/sbin/esxcfg-vswitch -M vmnic0 vSwitch0 -p VMotion
/usr/sbin/esxcfg-vswitch -M vmnic0 vSwitch0 -p LAB_VMs
/usr/sbin/esxcfg-vmknic -a -i 10.10.10.40 -n 255.255.255.0 VMotion
/usr/bin/vimsh -n -e “hostsvc/net/refresh”
/usr/bin/vimsh -n -e “internalsvc/refresh_network”

%post –interpreter=bash

cat < /etc/rc3.d/S99postconf
#!/bin/bash
/usr/sbin/esxcfg-firewall –openport 88,tcp,out,KerberosClientTCP
/usr/sbin/esxcfg-firewall –blockOutgoing
/usr/bin/vmware-vim-cmd hostsvc/vmotion/vnic_set vmk0
/usr/bin/vmware-vim-cmd hostsvc/memoryinfo 838860800
/usr/bin/vmware-vim-cmd vimsvc/auth/entity_permission_add vim.Folder:ha-folder-root lg-esxsu true Admin true
echo “Removing automated post script.”
rm /etc/rc3.d/S99postconf
EOF
chmod +x /etc/rc3.d/S99postconf

By: Mike La Spina

Mike La Spina — Wed, 03 Mar 2010 02:01:09 +0000

Kevin,

Have you tried a longer delay or added an additional delay after the last vswitch config.

Regards,
Mike

By: Kevin Brooksby

Kevin Brooksby — Wed, 03 Mar 2010 00:16:55 +0000

Ok please help me out here Mike as I am about to lose my marbles over this . I have been hacking away at this script for the better part of the day but every time I get to the post install section it sleeps for the designated 3 minute interval then seems to attempt to apply the post settings but always fails and eventually says something to the effect of “Failed to connect: Crypto Exception:error:02001002:system library:fopen:No such file or directory:/etc/vmware/ssl/rui.key. The virtual console f1 screen also spits out a ‘User-supplied script failed’ error message. It seems to apply my add group/users commands properly but not the network/svccon memory/vmotion checkbox tick commands. Strange thing is if I do ssh into the host post install and just drop the one liners in it applies without error. Any ideas?

%post –interpreter=bash
/bin/sleep 90
/usr/sbin/groupadd -g 5000 lg-esxsu
/usr/sbin/useradd -u 501 -G lg-esxsu super1
/usr/sbin/useradd -u 502 -G lg-esxsu super2
/usr/sbin/esxcfg-vswitch -A LAB_VMs vSwitch0
/usr/sbin/esxcfg-vswitch -A VMotion vSwitch0
/usr/sbin/esxcfg-vswitch -p VMotion -v 10 vSwitch0
/usr/sbin/esxcfg-vswitch -p LAB_VMs -v 11 vSwitch0
/usr/sbin/esxcfg-vswitch -M vmnic0 vSwitch0 -p VMotion
/usr/sbin/esxcfg-vswitch -M vmnic0 vSwitch0 -p LAB_VMs
/usr/sbin/esxcfg-vmknic -a -i 10.10.10.40 -n 255.255.255.0 VMotion
/usr/bin/vimsh -n -e “hostsvc/net/refresh”
/usr/bin/vimsh -n -e “internalsvc/refresh_network”
/usr/bin/vmware-vim-cmd hostsvc/memoryinfo 838860800

By: Manoj

Manoj — Fri, 19 Feb 2010 21:17:23 +0000

Hi Mike,

Ignore my previous posts on this topic and I did some extra checking and found this is an issue with the Starwind iSCSI implementation that I have. For some reason the even after mapping the datastores manually on reboot the mappings are lost yet the ESX iscsi initiator successfully maps the targets.

I will have to revert to running something like openfiler. I am currently testing my build process before I introduce it into our non-production system which has both SAN and iSCSI storage LUNS.

One more thing on the ks.cfg script I wanted to know what I might be doing wrong for setting the console memory to 800MB. I am using the below:

mv -f /etc/vmware/esx.conf /etc/vmware/esx.conf.old
sed -e ’s/boot\/memSize = \”300\”/boot\/memSize = \”800\”/g’ /etc/vmware/esx.conf.old >> /etc/vmware/esx.conf

Which used to work for me on ESX3 but in ESX4 for some reason it shows as 300.00MB (800.0 MB on restart) I have restarted the host 3 times but the change is not effected.

Thanks,
Manoj