5

Automatically creating volume partitions on system reboot

view full story
linux-howto

http://serverfault.com – I followed this guide: http://www.mongodb.org/display/DOCS/Amazon+EC2+Quickstart to set up Mongodb. It had me create a RAID 10 array out of the four devices on EBS. Then it had me create a physical volume, a volume group, and three logical volumes out of that RAID 10 array. Lastly it had me create ext4 filesystems out of the logical volumes and mount them. Now the quickstart guide had me put two things in place so that these steps would be replicated on reboot of the system. It had me add some instructions to the mdadm.conf file to automatically create the RAID 10 array, and it also had me (HowTos)

u85500's picture
Submitted by u85500 on 10/31/2012 – Made popular on 10/31/2012

Stories similar to Automatically creating volume partitions on system reboot

Basing this question on information I got in Why does df shows only half the size of a RAID10 array?, how should I create a RAID10 array using EBS volumes on Amazon, in a way that will not be redundant too much ?

As the linked question shows, I was using a guide from the 10gen company (that created mongodb) for how to create a raid10 array using EBS.

I encounter an issue with the size of the cre

Basing this question on information I got in Why does df shows only half the size of a RAID10 array?, how should I create a RAID10 array using EBS volumes on Amazon, in a way that will not be redundant too much ?

As the linked question shows, I was using a guide from the 10gen company (that created mongodb) for how to create a raid10 array using EBS.

I encounter an issue with the size of the cre

After a restart which stalled, I did a hard reboot and now mdadm can't see the RAID 6 array anymore, LVM can't find the volume which was on that raid array.

Is it possible to create a volume group from a logical volume instead of a physical volume? If so, are there any pitfalls in doing so?

Use case:

Installing OpenStack Compute on a a system that already has all of the physical volumes assigned to a singe volume group. The nova-volume service requires a separate volume group, as described in the documentation.

Before my installation of 12.04 server I read a lot on the forum about raid install and problems with Grub not being able to boot from anything other than a raid 1 array. However I had no problem at all with installation on a raid 6 array (5 disks) using mdadm.

What I did using Alternate 12.04 CD:

1. Choose Manual partion 2. Wipe all 5 drives 3. Assign all drives to a Raid 6 array 4.

I have installed 12.04 LTS on a machine that has 5 hard drives. I used the alternate CD to set up the 2 - 500 GB drives as the system RAID1 with LVM's (a system logical volume and swap logical volume). All works well, I've reboot the system and it boots fine.

So I have a system (a large instance running ubuntu-11.10 on Amazon EC2) where I've done something like

mdadm --create bigdata --level 0 --raid-devices=2 /dev/xvdb /dev/xvdc

I get a RAID device of two disks (ephemeral instance stores) at /dev/md/bigdata. Yay. reboot. I now have a RAID device of two disks at ip-10-811-498-737:bigdata. (ip anonymized).

This is kind of obnoxious.

I installed MongoDB in an Ubuntu Server VM inside VMware using these instructions: http://docs.mongodb.org/manual/tutorial/install-mongodb-on-ubuntu/. Everything works fine, and I can access MongoDB at the regular port 27017

./mongo 172.16.150.131:27017

Now, I want to make a dummy replica set and another db instance for testing my app.

I'm setting up a new file server using the same storage configuration as my existing server, but the procedure is failing and i cannot figure out why. my goal is to create a TrueCrypt volume on top of a RAID 10 volume. however, when i start truecrypt -c, it destroys the RAID volume.