Building Out XenServer 6.5 with USB Boot and Software RAID 10
-
Is mkfs.xfs not included in XenServer 6.5?
with mkfs -help I get mkfs.ext2
-
No it is not there by default best option would be mkfs.ext3
-
options found by default
-
Well that sucks. Ext3? Ugh.
-
Yeah I'm thinking he adjusted the command to be
mkfs.ext3 /dev/md0And hit enter, just waiting for confirmation.
-
Script has been updated.
-
@scottalanmiller remeber to change the line that targets devices sd[a-d] to sd[b-e],
There is also a typo in the last line,
device-config:device=/dev/md10 should be /dev/md0 following that guide -
@Romo The typo's have been addressed.
-
@DustinB3403 said:
@Romo The typo's have been addressed.
Great, this way more people can follow the guide
-
We should probably make a topic on building an off-host ISO and backup target server to complement this guide here.
Since we've already gone full tilt with the USB mdadm RAID 10....
-
@DustinB3403 said:
We should probably make a topic on building an off-host ISO and backup target server to complement this guide here.
Since we've already gone full tilt with the USB mdadm RAID 10....
Agreed.
-
I'm build a XenServer right now on laptop hard drives, using this guide.
So it'll tie in well with Virtualize everything.
Scott would you start a new topic with the basics of "why everything should be virtual" reference this topic, and what the "second server" will be for.
-
I wonder if it's possible to put this into a executable script that could just be downloaded, and run.
-
@DustinB3403 said:
I wonder if it's possible to put this into a executable script that could just be downloaded, and run.
Yes it could, that's part of the goal. We want to be able to fully automate this. In the long run you will likely want a prompt that will ask which devices to add to the array and after that it could do everything on its own.
-
What kind of impact does log writing have on the USB? Is that worth a concern at all?
-
It could be a concern, especially if you are producing a ton of logs.
But with the second server, you could point your logs to that server.
So your compute servers don't ever have to deal with a full Dom0 because of logging.
The backup USB is also there in-case it gets burnt out.
-
@DustinB3403 said:
It could be a concern, especially if you are producing a ton of logs.
But with the second server, you could point your logs to that server.
So your compute servers don't ever have to deal with a full Dom0 because of logging.
The backup USB is also there in-case it gets burnt out.
I wonder if there is a way to make most/all of it run in RAM like with VMware?
-
There might be, but why? You want your RAM for your VM's.
Just use the built in redirection to push it to an off-host server or a dedicated folder.
-
@johnhooks said:
What kind of impact does log writing have on the USB? Is that worth a concern at all?
It wears it out, send that to Loggly or ELK.
-
@DustinB3403 said:
There might be, but why? You want your RAM for your VM's.
Just use the built in redirection to push it to an off-host server or a dedicated folder.
I was just thinking, it already takes up 1GB, if you could have it only take up a small amount and store it in RAM, it might be a help.
