My response got held up in moderation to the CBT list, so posting to ceph-users and sending a copy to you as well to ensure you get it.
--
Artie,
I'd just use ceph-disk unless you need a config that it doesn't support. Its a lot fewer commands, pre-tested, and works.
That said, I had to create some journal partitions outside of ceph-disk recently, and here's what I did:
sgdisk --new=1:0:+10000M --change-name=1:"ceph journal" --partition-guid=1:R --typecode=1:45b0969e-9b03-4f30-b4c6-b4b80ceff106 /dev/$device
If you need more than one journal partition on the same device, modify the above command by replacing 1: with 2:
For an OSD data partition, modify the partition size, set the change-name to "ceph data" and the typecode UUID to: 4fbd7e29-9d25-41b8-afd0-062c0ceff05d
The list of typecode UUIDs is easily viewable in the ceph-disk source, here:
I assume this will work with CBT since it worked with Ceph OSDs.
--David
On Wed, Oct 21, 2015 at 5:34 PM, Artie Ziff <artie.ziff@xxxxxxxxx> wrote:
My inquiry may be a fundamental Linux thing and/or requiring basic
Ceph guidance.
According to the CBT ReadMe -- https://github.com/ceph/cbt
------------
Currently CBT looks for specific partition labels in
/dev/disk/by-partlabel for the Ceph OSD data and journal partitions.
...each OSD host partitions should be specified with the following gpt labels:
osd-device-<num>-data
osd-device-<num>-journal
------------
Does this mean that a disk formatted with fdisk in MBR/DOS format
style should be changed to GPT?
I've been taking some advice from peers directing use of fdisk.
What is recommended disk prep tool and partition format (GPT/MBR)?
Or should I be using ceph-disk exclusively? (and be done with it! <grin>)
Also on the CBT ReadMe is a script that Users are encouraged to
inspect: mkpartmagna.sh
https://github.com/ceph/cbt/blob/master/tools/mkpartmagna.sh
The core task is iterating over items in directory /dev/disk/by-id.
==> However, my /dev/disk/by-id is not populated with items. <==
I realize this is a Linux thing... however I am not familiar with it.
When I google the topic I appears to be called persistent block device naming.
Does the parted command create the necessary labels that CBT requires?
Is there an extra step required to make the labels appear in /dev/disk/by-id
Are the Ceph udev rules related to this disk by-id naming?
And finally, are the Ceph udev rules a requirement for a proper
installation of Ceph?
And if you read this far... bonus question. :)
In this parted command,
parted -s -a optimal /dev/$DEV mkpart osd-device-$i-data $sp% $ep%"
What function/feature do the variables $sp and $ep, hold for us?
Or what may have been the author's intent?
BTW, although cross--posted, I tried to set a reply-to for CBT list
only. We see how it goes. Thanks in advance.
-az
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
