Re: administrative notification (related to esxthin.pm)
When executing perl code inside a provisioning module, when I got o use the mail utility function, how can I know to $to address that the admins have configured to be the vcl administrator e-mail address? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796 On Oct 20, 2009, at 8:46 AM, Aaron Peeler wrote: If you don't care about the message also going to the vcld.log file. You can call the utils mail routine from any of the other modules. mail($to,$subject,$mailstring, $from) $from is optional Aaron --On October 19, 2009 5:12:55 PM -0400 Brian Bouterse bmbou...@gmail.com wrote: As esxthin.pm manages the image library hosted on a NetApp storage system, the module monitors for important events, and would like to notify an administrator more pro-actively then logging it as a critical. Is there a VCL utility function, or existing example code which would be able to send an e-mail to the VCL administrator? This code would be a stop-gap until a more formal storage provisioning architecture is worked out in the future. Thanks, Brian Brian Bouterse NEXT Services 919.698.8796
XMLRPC Woes
I am trying to drive VCL via XMLRPC and am having a lot of trouble. I have correctly attached the X-User, X-Pass, and X-APIVERSION headers along with my HTTP transport, with the following values: X-User: ad...@local X-Pass: thecorrectpassword X-APIVERSION is correct I know most of it's working because the error message I get is: xmlrpclib.Fault: Fault 6: 'Unable to authenticate passed in X-User' Which line of code verifies this value? Any ideas why this isn't working? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796
Reservation Timeout Problems
Here's what I experience: I will make a reservation, it will load successfully, I will NOT click on the 'Connect' button, wait 10 minutes, and the reseration goes to Reservation has timed out although the computer it was loaded on shows that it is available under 'Manage Computers.' Basically I want to provision VM's but not connect into them. For the image I am reserving, I have gone through as set Check for logged in user: to No under Manage Images-Profiles-Edit-Advanced. How can I make these reservations not time out? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796
Multiple Reservations For the Same Image Form the Same User
I am trying to place multiple reservations for the same image from the same user ('admin') at the same time. When I try to place the reservation for the second image I receive the error: The time you requested overlaps with another reservation you currently have. You are only allowed to have a single reservation at any given time. Please select another time to use the application. If you are finished with an active reservation, click Current Reservations, then click the End button of your active reservation. Is there a way I can bypass this check? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796
administrative notification (related to esxthin.pm)
As esxthin.pm manages the image library hosted on a NetApp storage system, the module monitors for important events, and would like to notify an administrator more pro-actively then logging it as a critical. Is there a VCL utility function, or existing example code which would be able to send an e-mail to the VCL administrator? This code would be a stop-gap until a more formal storage provisioning architecture is worked out in the future. Thanks, Brian Brian Bouterse NEXT Services 919.698.8796
full perl path inside VCL module
While I'm working on the esxthin.pm provisioning module, I will be using a config file placed in the same directory as esxthin.pm. For example, in my installation this would be /opt/vcl/lib/VCL/Module/ Provisioning/esxthin.conf 1) I tried to open a file but it only works if I give it the entire path. Is there an easy way inside a VCL perl provisioning module to know the path of itself? 2) Is there an easier way to configure a username, password, and IP for esxthin.pm to use besides using the variable table? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796
Re: VCL scaleup limits
Thanks for the scalability info. I had one more question: does the vcld component benefit from multiple cores? In other words, is the vcld component multithreaded? Thanks! Brian Brian Bouterse NEXT Services 919.698.8796 On Sep 9, 2009, at 12:51 PM, Aaron Peeler wrote: Good question. It's hard to put an exact number on it because of the different types of resources that can be made available(vm,bare-metal,lab machines), but we should be able to get theoretically close. Correct - multiple management nodes is an important part of the scaling. So the first question is how many resources can a single management node support. Additional things to consider are how the nodes get provisioned and the usage profile(asynchronous or synchronous). Also this is assuming one has robust web server, database, networking, storage and management nodes. The network, storage and management nodes are a big factor. Provisioning options - bare-metal vs. hypervised vs. stand-alone lab machines *bare-metal - installing an image to disk using xCAT: typically we have experienced anywhere from 150-200 blades per management node for asynchronous use. *hypervisor using vmware free server vs ESX/i with either persistent vs non-persistent mode. VMware ESX with network datastores for the vms and running in non- persistent mode. One could probably get 1000+ vms per management node maybe more. Again this assumes fast access/networking to storage (10G ethernet or fibre), a nice storage array and running 20+ vms per esx server. vmware free server - typically 5-10 vms per server *stand-alone machines: at NCSU we also use traditional lab machines when the university labs close. This mode is only brokering remote access to nodes, therefore no loading is going on. One could probably get 700+ machines per management node. Usage Profile: asynchronous vs synchronous synchronous usage is the most demanding usage, block allocations/ provisioning for a class, workshop or some other event that needs many nodes at a single event. asynchronous usage - users independently request nodes at any given time. This is spread out over time and has a lower provisioning load on the management node/s. Since we(ncsu) don't have the infrastructure to confirm these numbers this is just an educated guess based on past experiences based on what we do have. I would feel comfortable saying that if using multiple management nodes and an ideal HW setup, beefy blades (multi-core, extended memory blades, high-end storage, fat-pipes, etc), vcl could conservatively support several thousands nodes. Aaron --On September 10, 2009 12:20:39 AM +1000 Sengor seng...@gmail.com wrote: Hi Brian, I'm not certain of the exact numerics, however I do believe support for multiple management nodes (vcld's) is an intentional scale-out approach. Perhaps some of the guys @ NCSU know this one, I believe their VCL instance is currently the largest one in production. On Wed, Sep 9, 2009 at 11:55 PM, Brian Bouterse bmbou...@gmail.com wrote: Hi All, I'm wondering what the scalability limits of VCL? What are the limits of the vcld component, and why do we believe they exist? Does the frontend have any scalability limits? I'm trying to figure out what the reasonable number of VMs a single vcld installation can concurrently support. What do you think and why? Best, Brian Brian Bouterse NEXT Services 919.698.8796 -- sengork Aaron Peeler OIT Advanced Computing College of Engineering-NCSU 919.513.4571 http://vcl.ncsu.edu
tester from my new e-mail
Brian Bouterse NEXT Services 919.698.8796
Re: Is the free license restricted in ESXi ?
Here is a bit of interesting evidence to help you get to the heart of the problem. I took your command and changed the variables to match a configuration in the lab, here is the command I used: ./vmcontrol.pl --server 152.14.17.230 --username root --password password_here --operation poweroff --vmname vdc This command successfully turns on a host when executed on a linux box wit viperl installed and running from the /usr/lib/vmware-viperl/apps/ vm/ directory. Here's another consideration, when I look at at the Licensed Features of my ESX 3i servers, I don't have any Add-Ons, but I notice that yours have VMware Consolidated Backup which is installed with the Evaluation License, when one first turns on an unlicensed ESX 3i box. When the license file from vmware.com is installed, it removes this Add-On capability. Yours is not removed, so this leads me to believe that your license configuration is not correct. Have you reproduced the problem on a different physical hypervisor? When licensing an ESX server, what exactly is your procedure? There is a prompt you are asked for during the licensing process, what is your response to that? In short, take another look at it, because that command works great on our end, and we're using commodity ESX 3i with a license from vmware.com Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On May 17, 2009, at 9:22 AM, 杨杰 wrote: There are two conditions: 1. The free license is installed correctly( which has been confirmed today). When checking the licensed features, i get the information as follows: Licensed Features Licensed Source Serial Number: [the SN got from the website] ESX Server Edition ESX Server Standalone Licensed for 2 CPUs NAS Usage iSCSI Usage Up to 4-way virtual SMP SAN Usage Add-Ons [Nothing here because no addons are installed] On the condition, the CMD i typed and the ERROR is: [r...@management vm]# ./vmcontrol.pl --server 192.168.4.202 -- username root --password xjtudlc --operation poweroff --vmname Redh5 Error in 'Redh5' under host esxi2.p12.edu.cn: VM 'Redh5' can't be powered off SOAP Fault: --- Fault string: fault.RestrictedVersion.summary 2. No license is installed: After the former operation, i deleted the license and checked the licensed features, i get the information as follows: Licensed Features Licensed Source Evaluation30 day(s) remaining ESX Server Edition ESX Server Standalone Licensed for 2 CPUs NAS Usage iSCSI Usage Up to 4-way virtual SMP SAN Usage Add-Ons VMware Consolidated Backup Licensed for 2 CPUs Then i started the vm with the VIClient and executed the same cmd through the remote command console. The operation is successful without any error or warnings. The target vm is also stopped as the cmd. We have checked the installation of the license and the error is indeed here. i dont know how to solve it. virtual machine 'Redh5' under host esxi2.p12.edu.cn powered off 2009/5/16 Brian Bouterse bmbou...@ncsu.edu The ESX 3i license given to you from the website, is that installed? You can check under Configuration - Licensed Features to make sure you have your stuff licensed for N CPUs. I've used vmcreate and vmcontrol operations with the ESX 3i and it worked fine for me. Could you provide the commands which are problematic? I don't see them in the e-mail thread below. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On May 15, 2009, at 11:11 PM, 杨杰 wrote: HI, I have one more question, which is a problem in our experiment. We are now using the esxi with free license. When trying to take the operations like vmcreate or vmcontrol, we receive an error command saying that we are trying a restricted operation. The cmds are as shown: Error in 'Redh5' under host esxi2.p12.edu.cn: VM 'Redh5' can't be powered off SOAP Fault: --- Fault string: fault.RestrictedVersion.summary The question has been submitted to the mailing list some days before. As the reply, we confirmed the reason on the both conditions of non- license and free-license, but the operations still succeeded without license and failed with a free license. Some discussions in the vmware official forum show that the free license is restricted from modification operations. Have you encountered the problem ? And is there any way to solve the problem ? -- Yang Jie(杨杰) Team of MRMSS, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: +8613468883723;029-82665263 EXT.24 MSN: xtyangjie2...@yahoo.com.cn E-mail: xtyang...@gmail.com Laurence J. Peter http://www.brainyquote.com/quotes/authors/l/laurence_j_peter.html - Originality is the fine art of remembering what you hear but forgetting where you heard it. -- Yang Jie(杨杰) Team
VCL Architecture Change Proposal
I propose exploring an alternative VCL architecture through an experimental code branch at Apache.org. This code would be housed in the svn at apache.org as a branch, and would be explored/developed in parallel with the current 2.X branch. The purpose of this branch is to explore a VCL architecture that allows the following: To allow VCL manage fluid resources (virtualized) more effectively Increase modularity by decoupling VCL components from a single, monolithic database via APIs Manage storage resources Create a database abstraction layer Increase image library confidentiality and integrity I've put a first-pass design document on the apache VCL wiki which outlines the architecture proposed with some level of detail. That document is located here: http://cwiki.apache.org/confluence/display/VCL/Experimental+VCL+Architecture+Document I'd like feedback of all kinds from the Apache VCL community to explore the architecture's merits and challenges at both the architectural and implementation levels. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Adding an image into VCL to test the esx.pm with
A good VM to test the esx.pm deployment functionality of VCL with is located here: http://www.rpath.org/project/lamp/build?id=13548 Place all the files into a folder inside the golden area of your VCL image library. To add the image into the VCL here are some instructions for a linux image: https://wiki.oscr.ncsu.edu/wiki/index.php/Creating_a_base_linux_ESX_image_for_VCL Here is a high level of what those instructions should help you accomplish: First, be sure to rename all files to the naming convention: esx3- YOURIMAGENAME-v0 Boot up the VM and make sure of the following give it the right number of nics (2) make all nics dhcp ensure the private nic gets a 10. address from VCL when it DHCP's and the public nic gets an IP from your dhcp system ensure that the root user on your VCL VM can ssh to the image's 10. address with no password (setup authorized keys) Before shutting down the VM in the library for the last time, remove the udev NIC rules at (/etc/udev/rules.d/70-persistent-net.rules) E-mail the vcl-dev list if you have any problems. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Re: Much Improved VCL Virtual Appliance (VCL 2.1 unofficial RC) Released
You can check on a VM as it is being provisioned by watching the vcld log. You can also view the hypervisor directly by using the VMware Virtual Infrastructure Client pointing at your hypervisor. For instructions on creating a base linux VM image which the VCL VM can deploy out see here: https://wiki.oscr.ncsu.edu/wiki/index.php/Creating_a_base_linux_ESX_image_for_VCL For windows instructions, see here: https://wiki.oscr.ncsu.edu/wiki/index.php/Creating_a_base_windows_ESX_image_for_VCL Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 20, 2009, at 4:03 AM, 乔木 wrote: Hi, Thank you for your release. We have setup the environment for your VCL 2.1 Virtual Applicance. How could we know whether the VCL VM provision is OK? I didn't see any imformation about how to load a VM to the Hypervisor and run it. Thanks 2009/4/9 Brian Bouterse bmbou...@ncsu.edu Some other folks in the community are specializing the bare metal installs through some form of integration with xcat 2.1.1. To keep my sanity, I am specializing in virtualization only VCL deployments and functionality. Who in the community is in charge of the bare metal integration moving forward? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 8, 2009, at 10:31 PM, 杨杰 wrote: Hi, Thank you for your greate help ! According to the email, this release does not support the deployment to the bare metal images. Then how could i make it ? Will you give some direction? Or is there any higher release version recently ? Thanks again ! 2009/4/9 Brian Bouterse bmbou...@ncsu.edu I've finished building the new version of the VCL VM appliance which is ready to deploy virtual machines onto ESX and ESX 3i hypervisors. This VM does not contain xcat and cannot deploy bare metal images or vmware-server images. The announcement is happening on this list because this VM appliance should lower the barrier to working with VCL within this community. Just to be clear, this VM is not being released by Apache. The major improvements include: The entire VM is only 3.1GB so it should be easy for folks to download The setup has been greatly simplified from the older and deprecated version hosted at the http://www.ibiblio.org/vclvm/ site The VM comes with the ESX/ESX 3i provisioning module pre- installed and ready to use To get started working with this VM, you will need: At least one ESX or ESX 3i hypervisor Some form of NFS storage which can also be SSH'd into. A good example of this is a simple Linux box acting as an NFS server. The link to the instructions (which include a VM download link) are located below. These instructions are on wiki ... if you use them, and find improvements, please update the documentation! https://wiki.oscr.ncsu.edu/wiki/index.php/The_VCL_2.1_Virtual_Appliance Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 -- Yang Jie(杨杰) Team of MRMSS, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: +8613468883723;029-82665263 EXT.24 MSN: xtyangjie2...@yahoo.com.cn E-mail: xtyang...@gmail.com -- Best wishes, 乔木 MOE KLINNS Lab and SKLMS Lab, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: 15991676983 E-mail: qiao...@gmail.com
Re: Questions about Distributing the VMware ESXi Server
No virtual center is not required in any way for ESX 3i. When you say you're installing it for a month, do you mean you've downloaded the ESX3i .iso installer from vmware, burned it to a CD, and am installing it from the CD? ESX 3i has to be installed manually today, VCL cannot deploy the ESX 3i hypervisor onto bare metal through any provisioning technique (xcat, esx, etc ...). xcat isn't well suited to managing ESX resources. This is why the ESX provisioning module (esx.pm) was written to use the VMware API to drive VMware resources (ESX and ESX 3i). Are you using the esx.pm module? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 15, 2009, at 10:51 PM, 杨杰 wrote: I have been trying to install a VMware ESXi Server to a bare metal machine for a month, but things are not going on well. I am using VCL1.6+xCAT1.3 now. How could i make it in the current environment ? In addition, i got from the vmware bbs that the virtual center is demanded to install the esxi automatically. Should i install the virtual center ? I got some information from the mail some days ago that the function of install esxi is not supported by xcat1.3 completely, and there would be a version of vcl which could support xcat2.1 recently. Is it right ? And when or where could i get it ? Thank you ! -- Yang Jie(杨杰) Team of MRMSS, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: +8613468883723;029-82665263 EXT.24 MSN: xtyangjie2...@yahoo.com.cn E-mail: xtyang...@gmail.com
Re: Questions about Distributing the VMware ESXi Server
NetApp has demonstrated at their Kilo-client lab in Raleigh, that ESX3i is PXE installable on a large scale (500) blades. I've asked them to send me any notes they can share with the community on how they made this work. Are you able to have VCL deploy VMs onto pre-installed ESX3i servers with the esx.pm module? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 16, 2009, at 10:00 AM, 杨杰 wrote: God. It sounds like that i had spent mutiple days on a complex task. We are also trying to drive esx3i to operate the VMs through VCL now. It is keeping pace with the provisioning of the esx3i. In addition, I have seen some information to install the esx3i automatically through pxe, but it has not been proved proper till now. I will have a try later and tell you folks if any achievement. The link is as follows: http://communities.vmware.com/docs/DOC-6824 you can also see the links in that site. Thank you so much. 2009/4/16 Brian Bouterse bmbou...@ncsu.edu No virtual center is not required in any way for ESX 3i. When you say you're installing it for a month, do you mean you've downloaded the ESX3i .iso installer from vmware, burned it to a CD, and am installing it from the CD? ESX 3i has to be installed manually today, VCL cannot deploy the ESX 3i hypervisor onto bare metal through any provisioning technique (xcat, esx, etc ...). xcat isn't well suited to managing ESX resources. This is why the ESX provisioning module (esx.pm) was written to use the VMware API to drive VMware resources (ESX and ESX 3i). Are you using the esx.pm module? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 15, 2009, at 10:51 PM, 杨杰 wrote: I have been trying to install a VMware ESXi Server to a bare metal machine for a month, but things are not going on well. I am using VCL1.6+xCAT1.3 now. How could i make it in the current environment ? In addition, i got from the vmware bbs that the virtual center is demanded to install the esxi automatically. Should i install the virtual center ? I got some information from the mail some days ago that the function of install esxi is not supported by xcat1.3 completely, and there would be a version of vcl which could support xcat2.1 recently. Is it right ? And when or where could i get it ? Thank you ! -- Yang Jie(杨杰) Team of MRMSS, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: +8613468883723;029-82665263 EXT.24 MSN: xtyangjie2...@yahoo.com.cn E-mail: xtyang...@gmail.com -- Yang Jie(杨杰) Team of MRMSS, Xi'an Jiaotong University Department of Computer Science and Technology, Xi’an Jiaotong University TEL: +8613468883723;029-82665263 EXT.24 MSN: xtyangjie2...@yahoo.com.cn E-mail: xtyang...@gmail.com
Re: xp image power on fail
I've updated the VCL-130 JIRA ticket with the following: The solution to be implemented in the esx.pm module is a check to see if a virtual machine's RAM = 0. If it is then return a friendly, but critical error message. This should solve the user confusion problem, and the provisioning engine doesn't try to compensate for misconfiguration in other areas of the system. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 9, 2009, at 3:52 PM, Aaron Peeler wrote: I sort of disagree, but can understand your view point. My viewpoint is from user complaints and tracing through logs as to why their image didn't load. So it was easier to sanity check the image creator(the user) input values instead of pointing out to users, why their image broke. Maybe the web front end for the image profile does not except '0' or allow below a certain value? I think some level of sanity checking is needed somewhere. -A --On April 9, 2009 3:25:14 PM -0400 Andrew Brown brow...@gmail.com wrote: I agree. Ideally, I like to keep systems as generalized as possible, and that means not doing anything unnecessary and not making any unnecessary assumptions. One such assumption being that there is some kind of minimum ram amount for images that VCL supports. A ram size of 0 is obviously an error, so I wouldn't be against a check for 0 values and failing in that case... but the system already does just that, it's just vmware that does the check, not VCL. -Andrew On Thu, Apr 9, 2009 at 3:17 PM, Brian Bouterse bmbou...@ncsu.edu wrote: If an installation's configuration (in this case the RAM metadata for an image) isn't properly setup, should the provisioning system compensate for that? I submit to the community that a provisioning engine should take the values handed to it, provision them, and that is all. If the values handed to the provisioning engine don't make sense (like a ram size of 0) then the provisioning engine should cause an error. What do you think? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 9, 2009, at 1:01 PM, Aaron Peeler wrote: Also that might be an improvement for the esx.pm to set the default to 512MB if it's less than 512. Created JIRA issue VCL-130 https://issues.apache.org/jira/browse/VCL-130 here's the snippet from the vmware.pm module. It's not perfect(doesn't account for assigning too much ram) but it might help: # check for memory settings my $dynamicmemvalue = 512; if (defined($vmclient_imageminram)) { # preform some sanity check if (($dynamicmemvalue $vmclient_imageminram) ($vmclient_imageminram $vmhost_RAM)) { $dynamicmemvalue = $vmclient_imageminram; notify($ERRORS{'OK'}, 0, setting memory to $dynamicmemvalue); } else { notify($ERRORS{'WARNING'}, 0, image memory value $vmclient_imageminram out of the expected range in host machine $vmhost_RAM setting to 512); } } --On April 9, 2009 12:41:33 PM -0400 Wayne Schildhauer wschi...@linux.vnet.ibm.com wrote: I found it. The image data in the database had 0 instead of 512. It seems like we looked at everything but the most obvious Wayne F. Schildhauer IBM Corporation Research Triangle Park, NC - Original Message - From: Wayne Schildhauer wschi...@linux.vnet.ibm.com To: vcl-dev vcl-dev@incubator.apache.org Sent: Wednesday, 08 April, 2009 18:22 Subject: xp image power on fail I am sorry for the naive question to come, but we figured out why our Windows XPs VMs are not powering on. In the deployed vmx file on ESXi, esx3-windowsxp-v0.vmx, memsize = 0: !/usr/bin/vmware config.version = 8 virtualHW.version = 4 memsize = 0 displayName = windowsxp-bl1 guestOS = other deleted remaining This causes VMware ESXi to panic the VM with an ASSERT failure. Our master configuration file shows it being 512 MB (memsize = 512), and the slots appear to be configured for 512 MB as well. I suspect that memsize is not getting initialized, rather than overwritten, but I cannot trace where the object that is being given to esx.pm is originally generated. Perhaps in the reservation? ThanksWayne Aaron Peeler OIT Advanced Computing College of Engineering-NCSU 919.513.4571 http://vcl.ncsu.edu Aaron Peeler OIT Advanced Computing College of Engineering-NCSU 919.513.4571 http://vcl.ncsu.edu
Re: vmprofile and vmtype tables in vcl.sql
Since ESX and ESX 3i both use the VMware API, VCL can name them different things in the vmtype table, and try to treat them differently, but it's a useless distinction. Since the API's are the same, the integration effort should not be duplicating work by treating them differently. Consider ESX 4.0; creating a distinction between ESX 4i and ESX 4 is similarly useless since the management API (and therefore the integration work) is the same for both of them. A useful distinction would be between ESX 3.5 and ESX 4 since the API does change between those versions. Does this make sense to folks? To answer the question asked about ESX3i kickstart installs, ESX3i cannot be installed via kickstart. The only reason ESX3 could be kickstarted is because the service console was RHEL based. ESX3i contains no RHEL, and therefore no kickstart opportunity. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 8, 2009, at 3:24 PM, Aaron Peeler wrote: doh - hit send too soon... I don't think I've seen the vmwareGSX name used in a while at the vmware sites, so we could just drop that entry. Just need to double check the code first to see if it matters. Speaking of redundancy in the vmtype table vcl.sql file, I'm confused about why there are independent entries for ESX and ESX 3i. The esx.pm provisioning module uses the VMware API (http://www.vmware.com/support/pubs/sdk_pubs.html) which is designed by VMware to be identical with ESX and ESX 3i. This way, folks like us don't have to implement custom code for one ESX vs ESX 3i. Because of this, does it even matter which vmprofile.vmtypeid was added to the vcl.sql file? What do you think? The reason is to separate ESX standard server and ESXi. So maybe ESX3 gets renamed to ESX standard Server. The vmware.pm module supports both vmware Free Server and ESX Standard Server using the 'vmware-cmd' cmd via ssh on the host server. ESX3i is only cmdline managed by the vmware-tools api - right? Ultimately I'm not sure how much the vmtype table will matter in the long-run. Since the computer.privisioingid is being set to which provisioning module to use. Given that VCL can't deploy ESX 3i (the free version), I propose VCL shouldn't mess with the 3i hypervisors until this can be revisited as a Can ESX3i be installed via kiskstart? If so then it can be installed using the xcat provisioning module, as long as the hardware is supported and one has xCAT setup on their management node. Aaron
Re: VCL Appliance
I'm releasing a 3GB (ESX provisioning only) VCL VM this afternoon. I'm finishing the instructions now before I post it. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 7, 2009, at 10:01 AM, Mark Gardner wrote: The http link is not broken, just horribly slow. At the current rate, it will take 3-4 days to download it. Does anyone have a faster copy? Mark On 4/6/09, Mark Gardner m...@vt.edu wrote: I have been unable to download the VCL appliance. The torrent appears to not have any seeders and the http link appears to be broken. How do I download it? -- Mark Gardner -- -- Mark Gardner --
Re: vmprofile and vmtype tables in vcl.sql
Yes, one of the two vmwareGSX and vmwarefreeserver entries in the vmtype table should be removed because they are redundant. Speaking of redundancy in the vmtype table vcl.sql file, I'm confused about why there are independent entries for ESX and ESX 3i. The esx.pm provisioning module uses the VMware API (http://www.vmware.com/support/pubs/sdk_pubs.html ) which is designed by VMware to be identical with ESX and ESX 3i. This way, folks like us don't have to implement custom code for one ESX vs ESX 3i. Because of this, does it even matter which vmprofile.vmtypeid was added to the vcl.sql file? What do you think? Given that VCL can't deploy ESX 3i (the free version), I propose VCL shouldn't mess with the 3i hypervisors until this can be revisited as a larger issue. As a fix, how about if we set the vmprofile.imageid to 4 (No Image) because of this? What do you think? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 8, 2009, at 11:46 AM, Josh Thompson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Brian's commit to the vcl.sql file highlighted an issue to me. The vmprofile table includes an imageid field. Even though we don't ship any images, we do ship entries in the image table to serve as examples. In our documentation, we need to have one of two things: 1) tell users how to create the images that correspond with the entries we ship in the image table or 2) tell users that after they've created their own VM host server images and added them to the image table, they need to update the vmprofile through the web interface to assign the correct image to the profile The vmtype table has both vmwareGSX and vmwarefreeserver, which are the same thing. Can we remove one of them? Brian: For the VMware ESX SAN vmprofile entry you added, you have vmtypeid set to 5 which is vmwareESX3. Should that have been 6 which is vmwareESXi instead? (and maybe vmwareESXi should be vmware ESX3i?). Also, you have imageid set to 9 which is VMware ESX 3.5 standard server. There is no entry in the image table for an ESX 3i server, but given the above issue, there's really no image that would correspond to an entry for it in the image table anyway; so, it doesn't really matter that much. Josh - -- - --- Josh Thompson Systems Programmer Virtual Computing Lab (VCL) North Carolina State University josh_thomp...@ncsu.edu 919-515-5323 my GPG/PGP key can be found at pgp.mit.edu -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFJ3MbLV/LQcNdtPQMRAusZAJoDyLJx68lrr0L5d7C7CoITH3jYoQCfYKKV kzNCRfCzSwHSMfIrMrxt0JU= =uVO6 -END PGP SIGNATURE-
Re: mysql user permissions bug?
I've resolved my issue, but I do think there is a bug here. It seems the SELECT, INSERT, UPDATE, DELETE permissions are all the management node's database user requires. However, even though there is a parameter in /etc/vcl/vcld.conf for LockerWrtUser, the value the management node will check in with is when: LockerWrtUser=vcl I am not sure why. -Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 6, 2009, at 12:21 PM, Brian Bouterse wrote: I am doing a fresh installation and starting with a database built from the most recent vcl.sql from the SVN. I've configured my frontend and management node, according to these instructions: https://svn.apache.org/repos/asf/incubator/vcl/trunk/INSTALLATION I've added my management node to the DB, but I couldn't get the vcld process to start cleanly without the following complaint from the log: SNIP- |2615| WARNING |2615| 2009-04-06 12:17:50|2615| utils.pm:get_management_node_info(7206)|zero rows were returned from database select statement: |2615|SELECT |2615|managementnode.*, |2615|predictivemodule.name AS predictive_name, |2615|predictivemodule.prettyname AS predictive_prettyname, |2615|predictivemodule.description AS predictive_description, |2615|predictivemodule.perlpackage AS predictive_perlpackage, |2615| state.name AS statename |2615|FROM |2615|managementnode, |2615|module predictivemodule, |2615| state |2615|WHERE |2615|managementnode.predictivemoduleid = predictivemodule.id |2615| AND managementnode.stateid = state.id |2615|AND |2615|managementnode.hostname like 'dhcp15.cnl.ncsu.edu%' |2615| ( 0) utils.pm, notify (line: 691) |2615| (-1) utils.pm, get_management_node_info (line: 7206) |2615| (-2) vcld, main (line: 127) /usr/lib/sendmail: option requires an argument -- f 2009-04-06 12:17:50|2615|utils.pm:mail(1301)|SUCCESS -- Sending mail To: , PROBLEM -- vcld |2615| CRITICAL |2615| 2009-04-06 12:17:50|2615|vcld:main(131)|unable to retrieve management node information from database |2615| ( 0) utils.pm, notify (line: 691) |2615| (-1) vcld, main (line: 131) SNIP- I believe this is default user permissions bug because when I give my vcl user in the database full permissions (All Privileges) instead of the recommended in the instructions (SELECT, INSERT, UPDATE, DELETE), it works. What are the right permissions for a frontend user? What are the right permissions for a backend user? Do the instructions need to be updated? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Re: mysql user permissions bug?
Even though LockerWrtUser is there to indicate which database user the management node should use to login to the db, the only value the daemon will start with is vcl. Has anyone ever gotten the vcld to start successfully and checkin with the database when using something other than 'vcl' (specified in vcld.conf as LockerWrtUser)? Is that clearer? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 6, 2009, at 12:48 PM, Aaron Peeler wrote: Could you re-phrase this part? I don't understand what issue your highlighting? The value assigned to 'LockerWrtUser=' is the variable used when vcld starts to make the database connection. Aaron --On April 6, 2009 12:35:26 PM -0400 Brian Bouterse bmbou...@ncsu.edu wrote: I've resolved my issue, but I do think there is a bug here. It seems the SELECT, INSERT, UPDATE, DELETE permissions are all the management node's database user requires. However, even though there is a parameter in /etc/vcl/vcld.conf for LockerWrtUser, the value the management node will check in with is when: LockerWrtUser=vcl I am not sure why. -Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Apr 6, 2009, at 12:21 PM, Brian Bouterse wrote: I am doing a fresh installation and starting with a database built from the most recent vcl.sql from the SVN. I've configured my frontend and management node, according to these instructions: https://svn.apache.org/repos/asf/incubator/vcl/trunk/INSTALLATION I've added my management node to the DB, but I couldn't get the vcld process to start cleanly without the following complaint from the log: SNIP- | 2615| WARNING | 2615| 2009-04-06 12:17:50|2615| utils.pm:get_management_node_info(7206)|zero rows were returned from database select statement: | 2615|SELECT | 2615|managementnode.*, | 2615|predictivemodule.name AS predictive_name, | 2615|predictivemodule.prettyname AS predictive_prettyname, | 2615|predictivemodule.description AS predictive_description, | 2615|predictivemodule.perlpackage AS predictive_perlpackage, | 2615| state.name AS statename | 2615|FROM | 2615|managementnode, | 2615|module predictivemodule, | 2615| state | 2615|WHERE | 2615|managementnode.predictivemoduleid = predictivemodule.id | 2615| AND managementnode.stateid = state.id | 2615|AND | 2615|managementnode.hostname like 'dhcp15.cnl.ncsu.edu%' | 2615| ( 0) utils.pm, notify (line: 691) | 2615| (-1) utils.pm, get_management_node_info (line: 7206) | 2615| (-2) vcld, main (line: 127) /usr/lib/sendmail: option requires an argument -- f 2009-04-06 12:17:50|2615|utils.pm:mail(1301)|SUCCESS -- Sending mail To: , PROBLEM -- vcld | 2615| CRITICAL | 2615| 2009-04-06 12:17:50|2615|vcld:main(131)|unable to retrieve management node information from database | 2615| ( 0) utils.pm, notify (line: 691) | 2615| (-1) vcld, main (line: 131) SNIP- I believe this is default user permissions bug because when I give my vcl user in the database full permissions (All Privileges) instead of the recommended in the instructions (SELECT, INSERT, UPDATE, DELETE), it works. What are the right permissions for a frontend user? What are the right permissions for a backend user? Do the instructions need to be updated? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 Aaron Peeler OIT Advanced Computing College of Engineering-NCSU 919.513.4571 http://vcl.ncsu.edu
ESX Provisioning Module Update (and a few concerns)
I wanted to let folks know the ESX provisioning module should be finished, or at least in a beta form. I have closed the JIRA ticket (VCL-29) corresponding to the creation of this module. Here is a review of the new functionality: Manage multiple ESX or ESX 3i hypervisors Deploys virtual machines onto these hypervisors Supports virtual machine capture routing to allow for updating a VM image, and the creation of new derivative images One last improvement will include refactoring the way our module gathers the private IP address of the VM that is being provisioned. The change include deprecating the watching of the arp table in favor of monitoring the dhcpd.leases file. While VCL now supports ESX/ESX 3i netboot based virtual machines, the VCL architecture presents a lot of real challenges for making this a scalable solution, and I'd like to identify and discuss a few points/ concerns here. VCL requires an entry in the computers table for each VM, and this entry needs to be tied to a vmhost. By hard selecting the virtual machine entries in the computer table (a VM slot) up front, the decision about where to place the next virtual machine isn't handled effectively. Each VM slot gets statically assigned physical characteristics at creation time. This very quickly creates a situation where there is space in the datacenter for a particular VM on one hypervisor or another, but VCL can't figure it out because the large RAM slots got used first for other images, and now the image in question can't find a slot to meet it's meta-data requirements. VCL will incorrectly report that there is not space in the infrastructure, when there really is. This is bad. Also, it makes the setup much more difficult since an average, modern blade can run 20 - 30 VMs, and if you manage an entire blade center you're manually creating 350 computer table entries. Each entry requires multiple, manual updates to the database. This is not a tractable solution, and needs to be addressed. As a possible solution (or part of one), one major moving part is the placement decision for a particular reservation (tantamount to which hypervisor this VM will be reserved on). Placement today in VCL is decided in the front-end without asking the hypervisors what their capabilities are of accepting the next VM (or cluster of VMs) in question. One possible way around this is to create a ESX placement controller module which determines where to place things. This module would be part of the VCL backend (although it could be called from the frontend). The ESX provisioning module authors each VMX file on the fly based on meta-data from the database, so it is already dynamic enough. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Changing the password for a user
I've created a few usern...@local users, and this user information is stored in the VCL database. I've assigned these usernames random passwords so that I can send them to their owners. I'd like their owners to be able to change this password form the web interface. Is this possible? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Changing the Reservation Length
What is the easiest way to allow for longer reservation times ... say one day or 30 days? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Re: Reservation Problem
I'm still suspicious of your DNS resolution and configuration. In addition to creating the DNS entries in your lab's DNS server, did you connect VCL to that DNS by creating an entry similar to the one below in the /etc/resolv.conf on your VCL management node? nameserver ip_of_your_DNS_here Can you connect with virtual center to the blade you're trying to provision to when referencing it strictly by it's hostname? If you want to remove DNS resolution from the list of possible culprits, you can create an entry in /etc/hosts of your VCL management node similar to the line below. I think that should work. IP Address of esxib1esxibl1. Remember to restart networking (or the whole VCL management node) when testing these configuration changes. Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Mar 23, 2009, at 9:13 PM, Melba Lopez wrote: Ok, our lab already had DNS capabilities, so I've set esxibl1-3* *to be valid hostnames*. *However, it is still saying it can't find esxibl1. I've pinged that hostname from another machine on the same network and it works just fine. Also because it kept saying it couldn't load the VM, I've found another problem. The vmdk file successfully copies over into the inuse folder. When I manually add that VM to a blade, I'm noticing that the VM does not keep the network setings. I watched someone actually create this VM and I redid it myself, and the settings do not stay when its copied over. Not sure if it's related, but thought I might mention it. * 2009-03-24 04:28:57|24815|143:10|new|esx.pm:load(300)|wrote vmxarray to /mnt/vcl/inuse/esx3-lamp-v0-bl1/esx3-lamp-v0.vmx 2009-03-24 04:28:59|24815|143:10|new|vcld:REAPER(770)|setting $? to 0, exit status: 0 2009-03-24 04:28:59|24815|143:10|new|esx.pm:load(321)|Registered: |24815|143:10|new| No host found with name esxibl1. 2009-03-24 04:28:59|24815|143:10|new|esx.pm:load(330)|Power on command: /usr/lib/vmware-viperl/apps/vm/vmcontrol.pl --server 'esxibl1' -- vmname esx3-lamp-v0-bl1 --operation poweron --username vcl --password 'd834jfhd734jgfhf7' 2009-03-24 04:29:00|24815|143:10|new|vcld:REAPER(770)|setting $? to 0, exit status: 0 2009-03-24 04:29:00|24815|143:10|new|esx.pm:load(333)|Powered on: Virtual Machine esx3-lamp-v0-bl1 not found. 2009-03-24 04:29:01|24815|143:10|new|vcld:REAPER(770)|setting $? to 0, exit status: 0 2009-03-24 04:29:01|24815|143:10|new|vcld:REAPER(770)|setting $? to 0, exit status: 0 2009-03-24 04:29:01|24815|143:10|new|utils.pm:mail(1281)|SUCCESS -- Sending mail To: r...@localhost, PROBLEM -- esx.pm |24815|143:10|new| CRITICAL |24815|143:10|new| 2009-03-24 04:29:01|24815|143:10|new| esx.pm:load(341)|Could not query for VM in VI PERL API |24815|143:10|new| ( 0) utils.pm, notify (line: 683) |24815|143:10|new| (-1) esx.pm, load (line: 341) |24815|143:10|new| (-2) new.pm, reload_image (line: 664) |24815|143:10|new| (-3) new.pm, process (line: 266) |24815|143:10|new| (-4) vcld, make_new_child (line: 606) |24815|143:10|new| (-5) vcld, main (line: 353) |24815|143:10|new| WARNING |24815|143:10|new| 2009-03-24 04:29:01|24815|143:10|new|new.pm:reload_image(669)|esx3-lamp-v0 failed to load on esx3-lamp-v0-bl1, returning |24815|143:10|new| ( 0) utils.pm, notify (line: 683)* On Thu, Mar 19, 2009 at 7:12 PM, Josh Thompson josh_thomp...@ncsu.eduwrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Thu March 19 2009 5:45:40 pm Brian Bouterse wrote: always fake the DNS resolution by placing an entry /etc/resolv.conf That would be /etc/hosts :) Josh - -- - --- Josh Thompson Systems Programmer Virtual Computing Lab (VCL) North Carolina State University josh_thomp...@ncsu.edu 919-515-5323 my GPG/PGP key can be found at www.keyserver.net -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.9 (GNU/Linux) iD8DBQFJwt9xV/LQcNdtPQMRAiekAJ9Sna6on6gIZC8fO/1wKWFw9W3zIQCffxLj nKgBCGnhdhToNQ0EtzLhRC8= =Yjph -END PGP SIGNATURE-
VCL Web Front-end Troubleshooting
I installed and configured the VCL web front-end component with the VCL database component inside a VM, but I'm having trouble getting past the login screen. The general description is that I am presented with the standard VCL login screen, I enter the default user/pass, but it sends me right back to the login screen. There are no errors reported, but I can't get past the login screen. Any ideas of what is causing this, and how I can dig more into this problem to troubleshoot it? Thanks, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Trying to access vmhost_profile_datastore_path
I am trying to use the following line in the esx.pm module to get access to the datastore_path field of the 'vmhost' table. $self-data-vmhost_profile_datastore_path() The following error is thrown by the above line (168) in vcld.log: Uncaught exception from user code: OIO::Method error: Can't locate object method vmhost_profile_datastore_path via package VCL::DataStructure Package: VCL::Module::Provisioning::esx File: /opt/vcl/bin/../lib/VCL/Module/Provisioning/esx.pm Line: 168 Did this variable get pulled through the VCL::DataStructure.pm module appropriately? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796
Re: Trying to access vmhost_profile_datastore_path
That was exactly the problem; thanks! Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Mar 13, 2009, at 4:07 PM, Aaron Peeler wrote: ah - need the 'get_' Try $self-data-get_vmhost_profile_datastore_path() Aaron --On March 13, 2009 3:54:10 PM -0400 Brian Bouterse bmbou...@ncsu.edu wrote: I am trying to use the following line in the esx.pm module to get access to the datastore_path field of the 'vmhost' table. $self-data-vmhost_profile_datastore_path() The following error is thrown by the above line (168) in vcld.log: Uncaught exception from user code: OIO::Method error: Can't locate object method vmhost_profile_datastore_path via package VCL::DataStructure Package: VCL::Module::Provisioning::esx File: /opt/vcl/bin/../lib/VCL/Module/Provisioning/esx.pm Line: 168 Did this variable get pulled through the VCL::DataStructure.pm module appropriately? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 Aaron Peeler OIT Advanced Computing College of Engineering-NCSU 919.513.4571 http://vcl.ncsu.edu
VCL Database Abstraction Layer
Does the VCL architecture contains a database abstraction layer in its architecture? I'm curious what it would take to make VCL generalized to work with multiple database backends other than mySQL. Brian Brian Bouterse Secure Open Systems Initiative
Re: VCL, is this name kosher?
Within the world of apache nothing. Do all acronyms drip with meaning in Apache or do some just pick a few letters? There hasn't been a lot of discussion on this topic, yet it needs resolution. When does this get put to vote? Best, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Feb 18, 2009, at 11:12 PM, Alan D. Cabrera wrote: What does VCL stand for then? Regards, Alan On Feb 17, 2009, at 7:35 AM, Brian Bouterse wrote: I wanted to revisit this thread, because no conclusion was reached. The current status is that after much discussion on the list, no name changes were proposed, or ratified. There are ASF naming guidelines located here. Personally, I interpret this as a non-issue because the NCSU service is called the Virtual Computing Lab, and the ASF project name is VCL which is not the same. What do you think? my 2 cents, Brian Brian Bouterse Secure Open Systems Initiative 919.698.8796 On Jan 6, 2009, at 11:18 AM, Josh Thompson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Maybe we should look at this from a different angle - is there anything that *can* happen for the name to continue as VCL at ASF? Josh On Tuesday January 06, 2009, Alan D. Cabrera wrote: Thanks for your comments Matt. Well, it's after the holidays. Time to pick this thread back up. This issue needs to be resolved before we graduate. Regards, Alan On Dec 19, 2008, at 1:01 PM, Matt Hogstrom wrote: Alan, Thanks for starting the discussion about this. I think there are a few different issues so I'll try to separate them. The issues we need to address is the trademark of VCL and the other is the identity of the community. Perhaps we should talk about these in reverse order. First is the identity of the community. The move of the development for the VCL codebase from NC State to Apache marked the formation of a community outside of NC State. So, what NC State decides to do or not to do is not particularly relevant to the community collaborating at Apache. Given this separation, perhaps a different name is appropriate so their is a clear distinction between NC State's use of VCL and the Apache Incubator's community. As far as the name goes. Since Virtualization and hosting is, well, virtual, perhaps a name like spectre or Casper would be good :) (ya, I already can see the naming police coming). On Dec 19, 2008, at 10:19 AM, Henry E Schaffer wrote: Alan writes: ... If VCL is a trademark owned by NCSU then they would have to assign the trademark to the ASF or we would have to change our project's name. OK - I think we've reached a very clear question to be resolved. I think here's what needs to be resolved (at least in my mind): If VCL is a trademark owned by NCSU, and If NCSU assigns the ownership to the ASF - then would ASF give NCSU permission to continue its web site as is (possibly with an acknowledgement to the ASF)? Alternatively: If VCL is a trademark owned by NCSU, and If NCSU doesn't assign the ownership to the ASF, but gives permission to ASF to use VCL as part of naming this ASF project - could, and would the ASF use AVCL (Apache VCL) as the name? (Or some other formulation which would be an Apache name but clearly portray to users and potential users that it's the same project.) Does the above make sense? -- --henry schaffer - -- - --- Josh Thompson Systems Programmer Virtual Computing Lab (VCL) North Carolina State University josh_thomp...@ncsu.edu 919-515-5323 my GPG/PGP key can be found at pgp.mit.edu -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFJY4RhV/LQcNdtPQMRAtNJAJ9CViggEk5XIPl5DH/Q+cL2JerOoQCbBHsv O7WjXdHQWfQ6B3uUTkTr9+g= =iwnz -END PGP SIGNATURE-