possibly missing foreign key constraint and revisit old issues
Couple things: (1) The vmprofileid in vmhost table seems to point to vmprofile.id; I could be wrong. There does not seem to be a foreign key constraint there; I could delete vmprofile.id that is still being referenced. (2) Any updates from an old issue: http://www.mail-archive.com/vcl-dev@incubator.apache.org/msg00900.html The target is ESXi 3.5 host. Thanks
ESXi host, local disk
Replying to my own thread: >> (2) Any updates from an old issue: >>http://www.mail-archive.com/vcl-dev@incubator.apache.org/msg00900.html >> The target is ESXi 3.5 host. To get familiarity of the software, I've been using ESXi 3.5 local disk storage; so I specify the datastore to be ESXi's IP:/PATH. Hence, during image capture phase, the ssh commands seem to be going off to the ESXi host. The command seems to succeed all the time; however the return code is mostly 255 and sometimes 0. It's weird because I'm certain that commands being sent were successful; I reproduced it manually the weird behavior ( similar to http://www.mail-archive.com/vcl-dev@incubator.apache.org/msg00900.html ) I take that this is an sshd oddity on ESXi host. I saw entries of "localdisk" or "networkdisk" on vmprofile. Not sure how they are used. Is the setup of ESXi with local disk considered to be a supported option ? If so, how would I get around the sshd oddity for the ESXi's IP:/DATASTOREPATH issues during image capture ? Or I should do things differently for localdisk. I'll switch to NFS so that I can move forward. I wish to get some understanding on potentially what I did wrong or what I did was simply an "unsupported" setup. Thanks again.
VCL-2.1 SQL Constraints
Hi Devs, The vmprofileid in vmhost table seems to point to vmprofile.id; I could be wrong. There does not seem to be a foreign key constraint there; I could delete vmprofile.id that is still being referenced. Thanks
Re: ESXi host, local disk
mand(6180)|executing SSH command on 192.168.88.4: |13095|1:10|image| /usr/bin/ssh -i /etc/vcl/vcl.key -l root -p 22 -x 192.168.88.4 'cp /export/nfsstore/inuse/vmwarelinux-base10-v1/vmwarelinux-base10-v1-flat.vmdk /export/nfsstore/golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1-flat.vmdk' 2>&1 2010-06-01 13:52:35|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:52:35 2010-06-01 13:52:40|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:52:40 2010-06-01 13:52:45|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:52:45 2010-06-01 13:52:50|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:52:50 2010-06-01 13:52:55|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:52:55 2010-06-01 13:53:00|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:00 2010-06-01 13:53:05|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:05 2010-06-01 13:53:10|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:10 2010-06-01 13:53:15|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:15 2010-06-01 13:53:20|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:20 2010-06-01 13:53:25|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:25 2010-06-01 13:53:30|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:30 2010-06-01 13:53:35|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:35 2010-06-01 13:53:40|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:40 2010-06-01 13:53:45|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:45 2010-06-01 13:53:50|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:50 2010-06-01 13:53:55|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:53:55 2010-06-01 13:54:00|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:00 2010-06-01 13:54:05|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:05 2010-06-01 13:54:11|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:10 2010-06-01 13:54:16|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:16 2010-06-01 13:54:21|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:21 2010-06-01 13:54:26|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:26 2010-06-01 13:54:31|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:31 2010-06-01 13:54:36|13091|vcld:main(165)|lastcheckin time updated for management node 1: 2010-06-01 13:54:36 On Sat, May 29, 2010 at 2:34 PM, Ryan Johnson wrote: > SSH is disabled by default on the ESXi host and is 'unsupported' by Vmware. > I've noticed this behavior as well, where the command succeeds but it is > not reflected in the return code. You can enable SSH by going to the ESXi > console. If you go the NFS route make sure to use the address of the NFS on > the exporting computer rather than address of the mount of the ESXi host or > you will run into the same problems if SSH is not enabled. The idea that > the VCL developers have expressed is that using an NFS setup distributes the > workload among multiple drives which improves performance. > > On Sat, May 29, 2010 at 2:21 PM, My LinuxHAList > wrote: > >> Replying to my own thread: >> >> >> (2) Any updates from an old issue: >> >> >> http://www.mail-archive.com/vcl-dev@incubator.apache.org/msg00900.html >> >> The target is ESXi 3.5 host. >> >> To get familiarity of the software, I've been using ESXi 3.5 local >> disk storage; so I specify the datastore to be ESXi's IP:/PATH. >> >> Hence, during image capture phase, the ssh commands seem to be going >> off to the ESXi host. >> The command seems to succeed all the time; however the return code is >> mostly 255 and sometimes 0. >> It's weird because I'm certain that commands being sent were >> successful; I reproduced it manually the weird behavior >> ( similar to >> http://www.mail-archive.com/vcl-dev@incubator.apache.org/msg00900.html >> ) >> >> I take that this is an sshd oddity on ESXi host. >> >> I saw entries of "localdisk" or "networkdisk" on vmprofile. Not sure >> how they are used. >> >> Is the setup of ESXi with local disk considered to be a supported option ? >> If so, how would I get around the sshd oddity for the ESXi's >> IP:/DATASTOREPATH issues during image capture ? Or I should do things >> differently for localdisk. >> >> I'll switch to NFS so that I can move forward. >> >> I wish to get some understanding on potentially what I did wrong or >> what I did was simply an "unsupported" setup. >> >> Thanks again. >> >
Re: ESXi host, local disk
Ryan, thanks for the input and helpful discussion. I had a similar discussion with another helpful user on the topic. I think the management node was going to ssh into the vm machine to set ethernet to dhcpd, configure the sshd, shut it down, de-register it from vmware, etc. My original intention is that if I've properly setup the image; set ethernet to dhcpd, etc., then this would obviate the needs for the management node to login (and thus safely ignore the warnings). By having done what the management node going to do and put it in the image, will I be able to skip the step of having the management node to login to the image ? The process did copy over: ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1.vmx ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1.vmdk ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1-flat.vmdk Thanks On Tue, Jun 1, 2010 at 6:04 PM, Ryan Johnson wrote: > I don't think you are quite there yet. VCL needs to be able to ssh into the > image to perform some configuration. I am pretty sure the VM you are trying > to capture an image of needs to be registered with the host and powered on. > From the looks of the error below, the VM may not have been registered with > the host. Or perhaps the Image you are trying to capture is a registered > VM, but its name does not match the virtual machine VCL is looking for. > > 2010-06-01 13:52:28|13095|1:10|image|esx.pm:capture(661)|Powered off: > Virtual Machine vmwarelinux-base10-v1 not found. > > I real issue again has to do with this error where VCL is unable to SSH into > the VM. > > |13095|1:10|image| WARNING > |13095|1:10|image| 2010-06-01 > 13:52:07|13095|1:10|image|utils.pm:run_ssh_command(6252)|attempt 1/3: > failed to execute SSH command on vmwarelinux-base10-v1: echo > vmwarelinux-base10-v1 > /root/currentimage.txt, exit status: 255, SSH > exits with the exit status of the remote command or with 255 if an > error occurred, output: > |13095|1:10|image| ssh output (echo vmwar...): ssh: > vmwarelinux-base10-v1: Name or service not known > |13095|1:10|image| ( 0) utils.pm, notify (line: 737) > |13095|1:10|image| (-1) utils.pm, run_ssh_command (line: 6252) > |13095|1:10|image| (-2) esx.pm, capture (line: 650) > |13095|1:10|image| (-3) image.pm, process (line: 162) > |13095|1:10|image| (-4) vcld, make_new_child (line: 596) > |13095|1:10|image| (-5) vcld, main (line: 341) > > On Tue, Jun 1, 2010 at 3:55 PM, My LinuxHAList wrote: > >> I went ahead to switch it to NFS (helpful users have suggested it to >> me as work-arounds). >> I hope that I'm done with my initial image (given the logs entries, I >> think I'm done with the initial image creation - please point out if >> I'm wrong). >> >> This message might be more appropriate to vcl-user. >> >> === >> OUTPUT for vcld run on 2010-06-01 13:50:35 >> === >> 2010-06-01 13:50:35|13091|vcld:main(116)|vcld environment variable set >> to 1 for this process >> 2010-06-01 13:50:35|13091|utils.pm:rename_vcld_process(7901)|renamed >> process to 'vcld vcld' >> 2010-06-01 13:50:35|13091|utils.pm: >> get_management_node_info(6815)|management >> node info retrieved from database for managementnode >> 2010-06-01 13:50:35|13091|vcld:main(127)|retrieved management node >> information from database >> 2010-06-01 13:50:35|13091|vcld:main(140)|management_node_id >> environment variable set: 1 >> 2010-06-01 13:50:35|13091|vcld:main(148)|management node checkin >> interval is 5 seconds >> 2010-06-01 13:50:35|13091|vcld:main(149)|vcld started on >> managementnode.local >> 2010-06-01 13:50:40|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:50:40 >> 2010-06-01 13:50:45|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:50:45 >> 2010-06-01 13:50:50|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:50:50 >> 2010-06-01 13:50:55|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:50:55 >> 2010-06-01 13:51:00|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:51:00 >> 2010-06-01 13:51:05|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:51:05 >> 2010-06-01 13:51:10|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:51:10 >> 2010-06-01 13:51:15|13091|vcld:main(165)|lastcheckin time updated for >> management node 1: 2010-06-01 13:51:15 >> 2010-06-01 13:51:20|1
Re: ESXi host, local disk
| | 4 | Vmware ESX SAN|5 | 9 | NULL | /vmfs/volumes/NetApp | /vmfs/volumes/storage1 | Intranet2 | MCNC Public| networkdisk | NULL | NULL | | 6 | MyESXiNode|6 | 4 | NULL | 192.168.88.4:/export/nfsstore| /vmfs/volumes/nfsstore | VM Network | Virtual Machine Network| localdisk | root | NOTMYPASSWORD | ++---+--+-+--+--++++-+--++ Thanks On Fri, Jun 4, 2010 at 2:12 PM, Andy Kurth wrote: > No, the management node must be able to login to the VM via SSH during > capture. If this isn't working, then SSH also won't be working when the > image is loaded and reservations will fail. The management node must be > able to SSH into the loaded image to configure several things such as adding > the user account, checking for a user connection, etc. > > It looks like SSH is working from the management node to the VM host. You'll > need to get it working to the VM before proceeding. Your log output shows > the VM computer name configured in the VCL database as > 'vmwarelinux-base10-v1' matching the image name. This is a bit odd. You > probably will want to come up with a VM naming scheme and change the name of > your VM computer first to avoid confusion. You should be able to do this > via Manage Computers -> Edit Computer Information. Change the name of the > vmwarelinux-base10-v1 computer to something like vm-1. > > The private IP address of the VM needs to be in /etc/hosts on the management > node so that the hostname vm-1 resolves to its private address. You should > be able to run "ping vm-1". > > Next, get SSH working from the management node to vm-1. The following > command should work: > ssh -i /etc/vcl/vcl.key -l root -p 22 -x vm-1 > > Once SSH is working, shut down the VM and save a copy of it for safety. > Power it back on and attempt the VCL image capture process again. > > Hope this helps, > Andy > > > My LinuxHAList wrote: >> >> Ryan, thanks for the input and helpful discussion. I had a similar >> discussion with another helpful user on the topic. >> >> I think the management node was going to ssh into the vm machine to >> set ethernet to dhcpd, configure the sshd, shut it down, de-register >> it from vmware, etc. >> >> My original intention is that if I've properly setup the image; set >> ethernet to dhcpd, etc., then this would obviate the needs for the >> management node to login (and thus safely ignore the warnings). >> >> By having done what the management node going to do and put it in the >> image, will I be able to skip the step of having the management node >> to login to the image ? >> >> The process did copy over: >> ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1.vmx >> ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1.vmdk >> ./golden/vmwarelinux-base10-v1/vmwarelinux-base10-v1-flat.vmdk >> >> Thanks >> >> On Tue, Jun 1, 2010 at 6:04 PM, Ryan Johnson >> wrote: >>> >>> I don't think you are quite there yet. VCL needs to be able to ssh into >>> the >>> image to perform some configuration. I am pretty sure the VM you are >>> trying >>> to capture an image of needs to be registered with the host and powered >>> on. >>> From the looks of the error below, the VM may not have been registered >>> with >>> the host. Or perhaps the Image you are trying to capture is a registered >>> VM, but its name does not match the virtual machine VCL is looking for. >>> >>> 2010-06-01 13:52:28|13095|1:10|image|esx.pm:capture(661)|Powered off: >>> Virtual Machine vmwarelinux-base10-v1 not found. >>> >>> I real issue again has to do with this error where VCL is unable to SSH >>> into >>> the VM. >>> >>> |13095|1:10|image| WARNING >>> |13095|1:10|image| 2010-06-01 >>> 13:52:07|13095|1:10|image|utils.pm:run_ssh_command(6252)|attempt 1/3: >>> failed to execute SSH command on vmwarelinux-base10-v1: echo >>> vmwarelinux-base10-v1 > /root/currentimage.txt, exit status: 255, SSH >>> exits with the exit status of the remote command or with 255 if an >>> error occurred, output: >>> |13095|1:10|image| ssh output (echo vmwar...): ssh: >>> vmwarelinux-base10-v1: Name or service not known >>&
showTimeTable function
Hi,
I have a ESXi host (computer.id=1); I have an image [ No Apps(Linux) ]
which was successfully captured (I think).
I tried to make Reservation using the "now" button, it says "Selection
currently not available"; then I clicked the "View Time Table", no
computers show up.
The ESXi host is in vmhostinuse state (which is state.id=20) -- which
ESXi host needs to be in.
I'm assuming that my ESXi host is supposed to show up on I hit the
"View Time Table" if I intend it to be able to run "No Apps (Linux)".
Is this true ?
With ESXi host in the state of vmhostinuse, I don't think it will ever
show up, because the state.id=20.
Here are the code snippets:
if($computer_platformids[$id] != $platformid ||
($computerData[$id]["stateid"] != 2 &&
$computerData[$id]["stateid"] != 3 &&
$computerData[$id]["stateid"] != 6 &&
$computerData[$id]["stateid"] != 8) ||
$computerData[$id]["ram"] <
$imageData[$imageid]["minram"] ||
$computerData[$id]["procnumber"] <
$imageData[$imageid]["minprocnumber"] ||
$computerData[$id]["procspeed"] <
$imageData[$imageid]["minprocspeed"] ||
$computerData[$id]["network"] <
$imageData[$imageid]["minnetwork"] ||
! in_array($id, $mappedcomputers)) {
continue;
}
I could have completely misunderstood what "View Time Table" should show.
I'm expecting the time tables to show ESXi host(s) that are able to
run "No Apps (Linux)".
If I'm wrong, what is "View Time Table" designed to show ?
Thanks
Re: showTimeTable function
Hi Mike, List,
Apparently, my understanding is still a bit off.
Since I've captured the image successfully "No Apps (Linux)", name:
vmwarelinux-base10-v1, with the appropriate files copied over to
$DATASTORE/golden/vmwarelinux-base10-v1/
I'm thinking that I'll be able to spawn as many "instance" of the "No
Apps (Linux)" as I configure the ESXi to handle.
That was the idea behind capturing an image right ? How would I be
able to achieve that ?
I create another computer called "vm-tospawn", populate the mac
addresses field and assign it to my only ESXi host.
Nothing shows up on the "View Time Table". What are the fields that
say "vm-tospawn" is associated with the "No Apps (Linux)" image I
captured ? If I populate the mac address fields, if I spawn 2 version
of it, wouldn't I cause conflicts on the Mac address ?
I still need some guidance on how to go from successful image capture
to launch multiple of them on ESXi host.
What am I missing ?
Thanks
On Tue, Jun 8, 2010 at 10:59 AM, Waldron, Michael H
wrote:
> The ESXi host won't directly run your image, you have to define virtual
> machines and assign them to the ESXi host, which will then spin these up to
> run the image.
>
> Basic steps:
> - Click Manage Computers, Edit Computer Information, Submit
> - Add
> - Fill in hostname, IP address, RAM, Proc Speed, Network Speed, select
> 'virtual machine' under Type, VMWare provisioning engine, and appropriate
> computer group.
> - Assign mac addresses for public and private network interfaces of the
> virtual machine.
> - Update the computer table in the database with this info for the virtual
> machine and assign the values to the fields eth0macaddress, eth1macaddress.
> - Click Virtual Hosts
> - Select your ESXi host
> - Click Configure Host
> - Select the virtual machine(s) in the Unassigned VMs column and click Add.
>
> Hopefully I didn't leave out any steps.
>
>
> Mike Waldron
> Systems Specialist
> ITS Research Computing
> University of North Carolina at Chapel Hill
> CB #3420, ITS Manning, Rm 2509
> 919-962-9778
>
>
> -Original Message-
> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
> Sent: Monday, June 07, 2010 11:54 PM
> To: vcl-u...@incubator.apache.org; vcl-dev@incubator.apache.org
> Cc: My LinuxHAList
> Subject: showTimeTable function
>
> Hi,
>
> I have a ESXi host (computer.id=1); I have an image [ No Apps(Linux) ]
> which was successfully captured (I think).
> I tried to make Reservation using the "now" button, it says "Selection
> currently not available"; then I clicked the "View Time Table", no
> computers show up.
>
> The ESXi host is in vmhostinuse state (which is state.id=20) -- which
> ESXi host needs to be in.
>
> I'm assuming that my ESXi host is supposed to show up on I hit the
> "View Time Table" if I intend it to be able to run "No Apps (Linux)".
> Is this true ?
>
> With ESXi host in the state of vmhostinuse, I don't think it will ever
> show up, because the state.id=20.
> Here are the code snippets:
> if($computer_platformids[$id] != $platformid ||
> ($computerData[$id]["stateid"] != 2 &&
> $computerData[$id]["stateid"] != 3 &&
> $computerData[$id]["stateid"] != 6 &&
> $computerData[$id]["stateid"] != 8) ||
> $computerData[$id]["ram"] <
> $imageData[$imageid]["minram"] ||
> $computerData[$id]["procnumber"] <
> $imageData[$imageid]["minprocnumber"] ||
> $computerData[$id]["procspeed"] <
> $imageData[$imageid]["minprocspeed"] ||
> $computerData[$id]["network"] <
> $imageData[$imageid]["minnetwork"] ||
> ! in_array($id, $mappedcomputers)) {
> continue;
> }
>
> I could have completely misunderstood what "View Time Table" should show.
> I'm expecting the time tables to show ESXi host(s) that are able to
> run "No Apps (Linux)".
>
> If I'm wrong, what is "View Time Table" designed to show ?
>
> Thanks
>
Re: showTimeTable function
workdisk |
+++-+-+-+-+
1 row in set (0.00 sec)
>
>
> Mike Waldron
> Systems Specialist
> ITS Research Computing
> University of North Carolina at Chapel Hill
> CB #3420, ITS Manning, Rm 2509
> 919-962-9778
>
>
> -Original Message-
> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
> Sent: Tuesday, June 08, 2010 12:15 PM
> To: vcl-dev@incubator.apache.org
> Subject: Re: showTimeTable function
>
> Hi Mike, List,
>
> Apparently, my understanding is still a bit off.
>
> Since I've captured the image successfully "No Apps (Linux)", name:
> vmwarelinux-base10-v1, with the appropriate files copied over to
> $DATASTORE/golden/vmwarelinux-base10-v1/
>
> I'm thinking that I'll be able to spawn as many "instance" of the "No
> Apps (Linux)" as I configure the ESXi to handle.
> That was the idea behind capturing an image right ? How would I be
> able to achieve that ?
>
> I create another computer called "vm-tospawn", populate the mac
> addresses field and assign it to my only ESXi host.
> Nothing shows up on the "View Time Table". What are the fields that
> say "vm-tospawn" is associated with the "No Apps (Linux)" image I
> captured ? If I populate the mac address fields, if I spawn 2 version
> of it, wouldn't I cause conflicts on the Mac address ?
>
> I still need some guidance on how to go from successful image capture
> to launch multiple of them on ESXi host.
>
> What am I missing ?
>
> Thanks
>
>
> On Tue, Jun 8, 2010 at 10:59 AM, Waldron, Michael H
> wrote:
>> The ESXi host won't directly run your image, you have to define virtual
>> machines and assign them to the ESXi host, which will then spin these up to
>> run the image.
>>
>> Basic steps:
>> - Click Manage Computers, Edit Computer Information, Submit
>> - Add
>> - Fill in hostname, IP address, RAM, Proc Speed, Network Speed, select
>> 'virtual machine' under Type, VMWare provisioning engine, and appropriate
>> computer group.
>> - Assign mac addresses for public and private network interfaces of the
>> virtual machine.
>> - Update the computer table in the database with this info for the virtual
>> machine and assign the values to the fields eth0macaddress, eth1macaddress.
>> - Click Virtual Hosts
>> - Select your ESXi host
>> - Click Configure Host
>> - Select the virtual machine(s) in the Unassigned VMs column and click Add.
>>
>> Hopefully I didn't leave out any steps.
>>
>>
>> Mike Waldron
>> Systems Specialist
>> ITS Research Computing
>> University of North Carolina at Chapel Hill
>> CB #3420, ITS Manning, Rm 2509
>> 919-962-9778
>>
>>
>> -Original Message-
>> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
>> Sent: Monday, June 07, 2010 11:54 PM
>> To: vcl-u...@incubator.apache.org; vcl-dev@incubator.apache.org
>> Cc: My LinuxHAList
>> Subject: showTimeTable function
>>
>> Hi,
>>
>> I have a ESXi host (computer.id=1); I have an image [ No Apps(Linux) ]
>> which was successfully captured (I think).
>> I tried to make Reservation using the "now" button, it says "Selection
>> currently not available"; then I clicked the "View Time Table", no
>> computers show up.
>>
>> The ESXi host is in vmhostinuse state (which is state.id=20) -- which
>> ESXi host needs to be in.
>>
>> I'm assuming that my ESXi host is supposed to show up on I hit the
>> "View Time Table" if I intend it to be able to run "No Apps (Linux)".
>> Is this true ?
>>
>> With ESXi host in the state of vmhostinuse, I don't think it will ever
>> show up, because the state.id=20.
>> Here are the code snippets:
>> if($computer_platformids[$id] != $platformid ||
>> ($computerData[$id]["stateid"] != 2 &&
>> $computerData[$id]["stateid"] != 3 &&
>> $computerData[$id]["stateid"] != 6 &&
>> $computerData[$id]["stateid"] != 8) ||
>> $computerData[$id]["ram"] <
>> $imageData[$imageid]["minram"] ||
>> $computerData[$id]["procnumber"] <
>> $imageData[$imageid]["minprocnumber"] ||
>> $computerData[$id]["procspeed"] <
>> $imageData[$imageid]["minprocspeed"] ||
>> $computerData[$id]["network"] <
>> $imageData[$imageid]["minnetwork"] ||
>> ! in_array($id, $mappedcomputers)) {
>> continue;
>> }
>>
>> I could have completely misunderstood what "View Time Table" should show.
>> I'm expecting the time tables to show ESXi host(s) that are able to
>> run "No Apps (Linux)".
>>
>> If I'm wrong, what is "View Time Table" designed to show ?
>>
>> Thanks
>>
>
Re: showTimeTable function
Hi Mike, List,
On Tue, Jun 8, 2010 at 3:49 PM, Waldron, Michael H
wrote:
> Your privilege screen attachments didn't make it to the list.
Ooops. Re-attached it. I made sure that things that can be added to
the Resources are added and marked available, etc.
Probably, the current setup is not clean; but it's the shot gun
approach to ensure that all things are available.
> Make sure for the Resources section on the Privileges screen that the
> computer groups and the image groups are listed as available.
>
> You shouldn't have to fill in the currentimageid, preferredimageid,
> nextimageid, imagerevisionid fields in the computer table.
OK.
> If you ever use "bare-metal" or "full blade" images in the future, as opposed
> to vmware images, you'll want to separate your computer and image group
> assignments. The reason being that "bare-metal" installs use a different
> provisioning engine (xCAT), bare-metal images can't be provisioned on vmware
> computers and vmware images can't be provisioned on bare-metal servers. In
> your case it doesn't matter since you only have vmware image.
> For example:
> allComputers,newimages,allImages used for bare-metal
> allVMComputers, newvmimages, allVMImages used for vmware
Thanks again for the explanation. All these labels are now more
meaningful to me now.
I see. So, in a cleanly-separated. My vm-spawn-1 belongs to only
newvmimages, allVMImages, allVMComputers.
In an attempt to make this work, I just make a lot of things belonging
to a whole lot of other things.
I'm still at this point getting "Selection not currently available".
How should I troubleshoot this further ?
>
>
> Mike Waldron
> Systems Specialist
> ITS Research Computing
> University of North Carolina at Chapel Hill
> CB #3420, ITS Manning, Rm 2509
> 919-962-9778
>
>
> -Original Message-
> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
> Sent: Tuesday, June 08, 2010 3:01 PM
> To: vcl-dev@incubator.apache.org
> Subject: Re: showTimeTable function
>
>> I don't know how the View Time Table plays into things, I've never looked at
>> that before, mine is currently blank.
>> Unless I'm not understanding the setup, you have to have as many computers
>> defined as you want to have instances of your image running. For each
>> computer you define you assign a unique mac address. To make a reservation
>> for the image, there has to be an available computer for the reservation to
>> run on. If you want to have 10 concurrent reservations, you have to have 10
>> computers available.
>>
>> The image also must be assigned to an image group, and that image group has
>> to be mapped to a computer group which contains the computers the image can
>> run on. There is good explanation of this at the Documentation link on the
>> VCL home page.
>
> Thanks. I've been trying to follow the documentation on the "Images
> and Privileges" documentation.
> The documentation did not mention about creation of each machine like
> you described; now I understand that it's needed.
>
> I added "vm-spawn-1"; with Mac address, etc.
>
> Do you mean the following grouping, etc. I've got (I'm logged in as
> user 'admin'):
> 1) "No Apps (Linux) - image.id=10" belongs to {allImages,allVMImages}
> 2) allImages belongs to {allComputers,newimages,newvmimages}
> 3) allVMImages belongs to {allComputers,newimages,newvmimages}
> 4) vmnode2.local (ESXi host) and vm-spawn-1 belong to
> {allComputers,newimages,newvmimages}
> 5) Attached are my Privileges screens.
>
> Do I need to set currentimageid, preferredimageid or nextimageid,
> imagerevisionid for vm-spawn-1 ?
>
> I still can't get reservation on "No Apps (Linux)", when I select
> "Now" it still says "Selection currently not available".
>
> Thanks
>
>
>
>
> mysql> select * from computer;
> ++-+-++++--+-+-+--++---+-+---+---+--+---+---+++---+-+---+---+--+--++--++--+
Re: showTimeTable function
privileges.
>
> select resourcegroup.name,resourcetype.name,type,privnode.name from
> resourcegroup,resourcetype,resourcepriv,privnode where
> resourcegroupid=resourcegroup.id and resourcetypeid=resourcetype.id and
> privnodeid=privnode.id;
>
> Mike Waldron
> Systems Specialist
> ITS Research Computing
> University of North Carolina at Chapel Hill
> CB #3420, ITS Manning, Rm 2509
> 919-962-9778
>
>
> -Original Message-
> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
> Sent: Tuesday, June 08, 2010 4:52 PM
> To: vcl-dev@incubator.apache.org
> Subject: Re: showTimeTable function
>
> Hi Mike, List,
>
> On Tue, Jun 8, 2010 at 3:49 PM, Waldron, Michael H
> wrote:
>> Your privilege screen attachments didn't make it to the list.
>
> Ooops. Re-attached it. I made sure that things that can be added to
> the Resources are added and marked available, etc.
> Probably, the current setup is not clean; but it's the shot gun
> approach to ensure that all things are available.
>
>> Make sure for the Resources section on the Privileges screen that the
>> computer groups and the image groups are listed as available.
>>
>> You shouldn't have to fill in the currentimageid, preferredimageid,
>> nextimageid, imagerevisionid fields in the computer table.
>
> OK.
>
>> If you ever use "bare-metal" or "full blade" images in the future, as
>> opposed to vmware images, you'll want to separate your computer and image
>> group assignments. The reason being that "bare-metal" installs use a
>> different provisioning engine (xCAT), bare-metal images can't be provisioned
>> on vmware computers and vmware images can't be provisioned on bare-metal
>> servers. In your case it doesn't matter since you only have vmware image.
>> For example:
>> allComputers,newimages,allImages used for bare-metal
>> allVMComputers, newvmimages, allVMImages used for vmware
>
> Thanks again for the explanation. All these labels are now more
> meaningful to me now.
>
> I see. So, in a cleanly-separated. My vm-spawn-1 belongs to only
> newvmimages, allVMImages, allVMComputers.
> In an attempt to make this work, I just make a lot of things belonging
> to a whole lot of other things.
>
> I'm still at this point getting "Selection not currently available".
>
> How should I troubleshoot this further ?
>
>>
>>
>> Mike Waldron
>> Systems Specialist
>> ITS Research Computing
>> University of North Carolina at Chapel Hill
>> CB #3420, ITS Manning, Rm 2509
>> 919-962-9778
>>
>>
>> -Original Message-
>> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com]
>> Sent: Tuesday, June 08, 2010 3:01 PM
>> To: vcl-dev@incubator.apache.org
>> Subject: Re: showTimeTable function
>>
>>> I don't know how the View Time Table plays into things, I've never looked
>>> at that before, mine is currently blank.
>>> Unless I'm not understanding the setup, you have to have as many computers
>>> defined as you want to have instances of your image running. For each
>>> computer you define you assign a unique mac address. To make a reservation
>>> for the image, there has to be an available computer for the reservation to
>>> run on. If you want to have 10 concurrent reservations, you have to have 10
>>> computers available.
>>>
>>> The image also must be assigned to an image group, and that image group has
>>> to be mapped to a computer group which contains the computers the image can
>>> run on. There is good explanation of this at the Documentation link on the
>>> VCL home page.
>>
>> Thanks. I've been trying to follow the documentation on the "Images
>> and Privileges" documentation.
>> The documentation did not mention about creation of each machine like
>> you described; now I understand that it's needed.
>>
>> I added "vm-spawn-1"; with Mac address, etc.
>>
>> Do you mean the following grouping, etc. I've got (I'm logged in as
>> user 'admin'):
>> 1) "No Apps (Linux) - image.id=10" belongs to {allImages,allVMImages}
>> 2) allImages belongs to {allComputers,newimages,newvmimages}
>> 3) allVMImages belongs to {allComputers,newimages,newvmimages}
>> 4) vmnode2.local (ESXi host) and vm-spawn-1 belong to
>> {allComputers,newimages,newvmimages}
>> 5) Attached are my Privileges screens.
>>
>> Do I need to set currentimageid, preferredimageid or nextimageid,
>>
Re: showTimeTable function
> Mike has already mentioned some of these items, but I'll list them all out > together to make it easier to check. > > - -put "No Apps(Linux)" in the allVMimages group It belong to allVMImages (and nothing else) > - -create a new computer group named "All VMs" > - -map the allVMimages image group to the "All VMs" computer group "allVMImags" maps to "AllVMs" and nothing else. > - -put vm-spawn-1 and any other VMs you've created in the "All VMs" computer > group vm-spawn-1 and vm-base10-v1 are grouped to AllVMs vmnode2.local (my ESXi host) is mapped to "allComputers" (and nothing else). > - -add "All VMs" to the VCL node on the privilege tree with the available > attribute Completed this one. > - -map the AllManagementNodes management node group to the "All VMs" computer > group allManagementNodes is mapped to "AllVMs" and nothing else. > - -make sure the computer specs for vm-spawn-1 are greater than or equal to > the > image specs for "No Apps(Linux)" They are equal. > - -make sure vm-spawn-1 is in the available state Yes. > - -make sure vm-spawn-1 is assigned to ESXi host in the Virtual Hosts part of > the site Yes. > - -make sure vcld is running on your management node and is regularly checking > in to the database Yes. > - -make sure your management node is in the AllManagementNodes group Yes. > - -make sure the schedule associated with vm-spawn-1 is available during the > time you've selected for your reservation (if you haven't messed with your > schedules, this shouldn't be an issue) Did not mess with this. I think it's 24x7. And now when I select "No Apps (Linux)" it says "Estimated Load Time < 5 Minutes". Thanks Mike, Josh, List. I see "My Reservation" eventually failed. It is a progress. Thanks > > I think I covered everything. Let us know what you find. > > Josh > > On Wednesday June 09, 2010, Waldron, Michael H wrote: >> Nothing obvious is jumping out at me here. I'm running out of ideas, >> perhaps someone else can chime in. >> >> The only other thing I can think of is to verify that your virtual machines >> are in the 'available' state. >> >> Mike Waldron >> Systems Specialist >> ITS Research Computing >> University of North Carolina at Chapel Hill >> CB #3420, ITS Manning, Rm 2509 >> 919-962-9778 >> >> >> -Original Message- >> From: My LinuxHAList [mailto:mylinuxhal...@gmail.com] >> Sent: Wednesday, June 09, 2010 9:41 AM >> To: vcl-dev@incubator.apache.org >> Subject: Re: showTimeTable function >> >> Thanks again Mike for helping with this. >> >> mysql> select resourcegroup.name,resourcetype.name,type,privnode.name >> from resourcegroup,resourcetype,resourcepriv,privnode where >> resourcegroupid=resourcegroup.id and resourcetypeid=resourcetype.id >> and privnodeid=privnode.id;; >> +++-+---+ >> >> | name | name | type | name | >> >> +++-+---+ >> >> | allComputers | computer | available | VCL | >> | allComputers | computer | administer | VCL | >> | allComputers | computer | manageGroup | VCL | >> | allComputers | computer | available | admin | >> | allComputers | computer | administer | admin | >> | allComputers | computer | manageGroup | admin | >> | allComputers | computer | available | newimages | >> | allComputers | computer | administer | newimages | >> | allComputers | computer | manageGroup | newimages | >> | newimages | computer | available | VCL | >> | newimages | computer | administer | VCL | >> | newimages | computer | manageGroup | VCL | >> | newimages | computer | available | admin | >> | newimages | computer | administer | admin | >> | newimages | computer | manageGroup | admin | >> | newimages | computer | available | newimages | >> | newimages | computer | administer | newimages | >> | newimages | computer | manageGroup | newimages | >> | newvmimages | computer | available | VCL | >> | newvmimages | computer | administer | VCL | >> | newvmima
Documentation Improvements on "Create a Linux Base Image"
Hi, If I have this understood right, I have a minor suggestion on the improvements on the page. https://cwiki.apache.org/confluence/display/VCL/Create+a+Linux+Base+Image The instructions need to be changed a bit for it to work for RedHat/CentOS 5. It has to do with the default sshd startup/stop script (as supplied by RedHat/CentOS 5) which does a "killproc $SSHD" on stop which will kill both internal and external daemons (even when invoked on the external sshd script). It's needed since the external sshd will be shutdown at some VCL processes and the internal sshd stilll needs to listen. For testing purposes, one can change it to kill on the PID file. To be ultra-safe, one can check the process name with the PID before killing. Thanks
minor bug fix
Hi Devs,
I'm somewhat certain that this is a bug and not entirely sure if this
has been reported:
/usr/local/vcl/lib/VCL/Module/Provisioning/esx.pm
Line 1019:if ($l =~ /([a-zA-Z1-9]*):\s*([-_.a-zA-Z1-9]*)/) {
should read if ($l =~ /([a-zA-Z0-9]*):\s*([-_.a-zA-Z0-9]*)/) {
Thanks
