I have been testing OperQRM 4.3 and I have found some problems that could be fixed.
In Addition I would like to see more info about the NetAPP plugin because I have been able to add iscsi lun to a Server.
When you create a volume , the plugin create a lun with 75% of the space of the volume, but it is not mapped to any initiator. then when you want to add an image the lun does not appear as root filesystem.
Any help on this....
Thanks
Carlos Gomez
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
many thanks for your feedback about the netapp plugin. We could not further progress with it because we do not have access to a netapp filer any more.
The lun not appearing may be caused by the storage-subsystem update we did with 4.3.
For the initiator group (as I remember from the development) we created a default initiator group for all our nodes since storage authentication was not yet implemented that time (that came with 4.3).
Any help on this is very much appreciated. Basically all commands are running via expect so if you could test and eventually provide the expect commands needed to e.g. authenticate and map the lun to an initiator group, create an initiator group etc. that would help a lot.
Not sure when we get access to a Netapp filer in the future.
many thanks in advance,
Matt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I can help with the netapp part . I see that you use expect with the netapp plugin so I can add the new stuff...however after authenticate and map the lun ..
how do you load the ISCSI initiator on the default kernel , so the appliance can use it has the root
image....
In addition I have an HBA card on a different nic for the resource that I want to provision. Can I use it? Any Ideas ?
I am using Centos 5.2 64bits as OPENQRM 4.3 node.
Thanks again....
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
:) great about your offer to help with the netapp plugin.
Please take a look at the iscsi-plugin (or lvm-storage plugin which comes with nfs, iscsi and aoe support).
For getting the iscsid/iscsiadm to the initrd and connect the lun please check :
/usr/lib/openqrm/plugins/iscsi-storage/web/root-mount.iscsi-deployment
-> this is being downloaded by a resource deployed with an iscsi-image.
For the authentication please look at :
/usr/lib/openqrm/plugins/iscsi-storage/web/openqrm-iscsi-deployment-auth-hook.php
-> the engine runs this when an appliance starts/stops
about the separated hba card:
You can use it but you may need to adapt the root-mount.deployment hook for you plugin to use the hba card to connect the lun instead of the regular nic.
hope it helps + have a nice day,
Matt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Can you give me a quick explanation of how manage the boot services?
I have used pxe boot and kickstart servers before and I noticed that OpenQRM set up the vmlinuz and initrd file. In addition, the pxelinux.cfg sends a openqrm parameter on the append entry that I do not understand. After this process how openqrm tells the system what to do with the root fs.
Thanks
Carlos Gomez
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
good, you found out a lot by yourself already.
The "openqrm" kernel commandline parameter is to let the booting server know where the openQRM server is (the ip). Then it sends a "node-parameters" request via http. The openQRM server answers this with the full configurtion parameter for the booting resource (resource-, image- ,appliance- and plugin/bootservice infos. After the server received this info it either stays in the "idle" stage or it deploys. When it deploys it handles over the responsiblity to mount the root-fs to a storage-plugin. Then it downloads the root-mount file of the storage plugins and runs the "mount_rootfs" function.
Then it basically just inits into the mounted rootfs.
hope it helps + have a nice day,
Matt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have been testing OperQRM 4.3 and I have found some problems that could be fixed.
In Addition I would like to see more info about the NetAPP plugin because I have been able to add iscsi lun to a Server.
When you create a volume , the plugin create a lun with 75% of the space of the volume, but it is not mapped to any initiator. then when you want to add an image the lun does not appear as root filesystem.
Any help on this....
Thanks
Carlos Gomez
Hi Carlos,
many thanks for your feedback about the netapp plugin. We could not further progress with it because we do not have access to a netapp filer any more.
The lun not appearing may be caused by the storage-subsystem update we did with 4.3.
For the initiator group (as I remember from the development) we created a default initiator group for all our nodes since storage authentication was not yet implemented that time (that came with 4.3).
Any help on this is very much appreciated. Basically all commands are running via expect so if you could test and eventually provide the expect commands needed to e.g. authenticate and map the lun to an initiator group, create an initiator group etc. that would help a lot.
Not sure when we get access to a Netapp filer in the future.
many thanks in advance,
Matt
Thanks Matt....
I can help with the netapp part . I see that you use expect with the netapp plugin so I can add the new stuff...however after authenticate and map the lun ..
how do you load the ISCSI initiator on the default kernel , so the appliance can use it has the root
image....
In addition I have an HBA card on a different nic for the resource that I want to provision. Can I use it? Any Ideas ?
I am using Centos 5.2 64bits as OPENQRM 4.3 node.
Thanks again....
Hi Carlos,
:) great about your offer to help with the netapp plugin.
Please take a look at the iscsi-plugin (or lvm-storage plugin which comes with nfs, iscsi and aoe support).
For getting the iscsid/iscsiadm to the initrd and connect the lun please check :
/usr/lib/openqrm/plugins/iscsi-storage/web/root-mount.iscsi-deployment
-> this is being downloaded by a resource deployed with an iscsi-image.
For the authentication please look at :
/usr/lib/openqrm/plugins/iscsi-storage/web/openqrm-iscsi-deployment-auth-hook.php
-> the engine runs this when an appliance starts/stops
about the separated hba card:
You can use it but you may need to adapt the root-mount.deployment hook for you plugin to use the hba card to connect the lun instead of the regular nic.
hope it helps + have a nice day,
Matt
Hi Matt,
Can you give me a quick explanation of how manage the boot services?
I have used pxe boot and kickstart servers before and I noticed that OpenQRM set up the vmlinuz and initrd file. In addition, the pxelinux.cfg sends a openqrm parameter on the append entry that I do not understand. After this process how openqrm tells the system what to do with the root fs.
Thanks
Carlos Gomez
Hi Carlos,
good, you found out a lot by yourself already.
The "openqrm" kernel commandline parameter is to let the booting server know where the openQRM server is (the ip). Then it sends a "node-parameters" request via http. The openQRM server answers this with the full configurtion parameter for the booting resource (resource-, image- ,appliance- and plugin/bootservice infos. After the server received this info it either stays in the "idle" stage or it deploys. When it deploys it handles over the responsiblity to mount the root-fs to a storage-plugin. Then it downloads the root-mount file of the storage plugins and runs the "mount_rootfs" function.
Then it basically just inits into the mounted rootfs.
hope it helps + have a nice day,
Matt