I have create an EBS drive, attached it to the Instance and created file system using mkfs.ext3.
Now i want to unmount and delete the drive, i've tried many things but nothing seems to work. Although i am able to detach the drive from instance and delete using EC-2 Console,
but when i am checking partition using df -hk it is still showing the drive.
[ec2-user#XXXXXXXXXXXXXX ~]$ df -hk
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/xvda1 8256952 1075740 7097356 14% /
tmpfs 304368 0 304368 0% /dev/shm
/dev/xvdf 30963708 176196 29214648 1% /media/newdrive
And more over when i try to use any other command like "fdisk -l" or and all or trying to browse the drive's folders, the putty session hangs.
I am new to EC2 cloud and also to Linux.
How about this?
You need to run as:
sudo umount /dev/xvdf
umount -dRf /media/newdrive
umount needs mountpoint not a devicetype like /dev/xvdf
Related
I have a server with the following disk structure:
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 219G 192G 17G 93% /
tmpfs 16G 0 16G 0% /lib/init/rw
udev 16G 124K 16G 1% /dev
tmpfs 16G 0 16G 0% /dev/shm
/dev/sda2 508M 38M 446M 8% /boot
/dev/sdb1 2.7T 130G 2.3T 5% /media/3TB1
I am interested in making backup of the whole server on my local machine. When the time comes I want to be able to restore a new server from my local machine backup. What procedure do you recommend?
I tried rsync, but the indexing took extremely long so I aborted it. Than I used scp, and well, it is currently working. There is lots of symbolic links that weren't transferred to the local machine, and I worry I won't be able to restore it later on.
Since your sda isn't very large and a lot of it is used anyway, I'd create a complete backup of the block device. Your sdb, however, is very large and used only to a small part. Of that I'd create a file system backup.
Boot your server with a Ubuntu live CD and become root (sudo su -).
Attach your backup medium (I assume it's mounted as /mnt/backup/ in the following).
Create a block device backup of sda: cat /dev/sda > /mnt/backup/sda
Mount your sdb (I assume it's mounted as /media/3TB1/ in the following).
Create a file system backup of sdb: rsync -av /media/3TB1/ /mnt/backup/sdb/
For restoring the backup later:
Boot your server with a Ubuntu live CD and become root (sudo su -).
Attach your backup medium (I assume it's mounted as /mnt/backup/ in the following).
Restore the block device backup of sda: cat /mnt/backup/sda > /dev/sda
Mount your sdb (I assume it's mounted as /media/3TB1/ in the following).
Restore the file system backup of sdb: rsync -av /mnt/backup/sdb/ /media/3TB1/
There are more fancy ways of doing it for sure. But this routine worked for me lots of times.
A backup of that size should take a long time to copy over the internet in any case: rsync, cp , dd ..etc, the time taken to copy the file depends on your internet speed.
In my opinion, rsync is the way to go, but if you're not willing to wait that long for the download to complete (I wouldn't either) I highly suggest backing your disk up on another remote server, unless you don't plan on restoring it later since uploading would be a pain too (especially on ADSL).
You have a few options:
Ask your data center for disk redundancy.
A cheap and highly unrecommended solution is to backup your most important data on a file sharing web service, eg. Dropbox (As far as I remember they had a shell API for many tasks including uploading files, which can be used for automatic backups).
Wait for the download to finish.
Go with #Alfe's solution, which is pretty neat in my opinion.
I don't have much experience with Linux and mounting/unmounting things. I'm using Amazon AWS, have booting up EC2 with Ubuntu image, and have attached a new EBS volume to the EC2. From the dashboard, I can see that the volume is attached to :/dev/sda1.
Now, I see from this guide from Amazon that the path will likely be changed by the kernel. So it's most likely that my /dev/sda1 device will be mounted on, maybe, /dev/xvda1.
So I logged in using terminal. I do ls /dev/ and I indeed see xvda1 on there. But I also see xvda. Now I want to format the device. But I don't know if the unformatted device is attached to xvda1 or xvda. I cannot list the content of /dev/xvda1 and /dev/xvda (it says ls: cannot access /dev/xvda1/: Not a directory). I guess I have to format it first.
I tried to format using sudo mkfs.ext4 /dev/xvda1. It says: /dev/xvda1 is mounted; will not make a filesystem here!.
I tried to format using sudo mkfs.ext4 /dev/xvda. It says: /dev/xvda is apparently in use by the system; will not make a filesystem here!
How can I format the volume?
EDIT:
The result of lsblk command:
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
xvda 202:0 0 8G 0 disk
`-xvda1 202:1 0 8G 0 part /
I then tried to use the command sudo mkfs -t ext4 /dev/xvda, but the same error message appears: /dev/xvda is apparently in use by the system; will not make a filesystem here!
When I tried to use the command mount /dev/xvda /webserver, error message appears: mount: /dev/xvda already mounted or /webserver busy. Some website indicate that this also probably because a corrupted or unformatted file system. So I guess I have to be able to format it first before able to mount it.
First of all you are trying to format /dev/xvda1, which is root device. Why ??
Second if you have added a new EBS, then follow below steps.
List Block Device's
This will give you list of block device attached to your EC2 which will look like
[ec2-user ~]$ lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
xvdf 202:80 0 100G 0 disk
xvda1 202:1 0 8G 0 disk /
Out of this xvda1 is the / (root) and xvdf is the one that you need to format and mount ( for the new EBS)
Format Device
sudo mkfs -t ext4 device_name # device_name is xvdf here
Create a Mount Point
sudo mkdir /mount_point
Mount the Volume
sudo mount device_name mount_point # here device_name is /dev/xvdf
Make an entry in /etc/fstab
device_name mount_point file_system_type fs_mntops fs_freq fs_passno
Execute
sudo mount -a
This will read your /etc/fstab file and if it's OK. it will mount the EBS to mount_point
I know if I put file on /dev/shm, then it is put on RAM of the server.
And if I put in my home directory, it is put on NFS.
And i know there is a command to tell if a given location on NFS or maybe RAM, what's that command?
Ex, how can I be sure my home directory is on NFS? I remember by using that command, some prints "NFS" can be seen
You can use the df command to show you the directory's mount point:
[mrsam#octopus ~]$ df -h .
Filesystem Size Used Avail Use% Mounted on
/dev/md0 178G 32G 137G 19% /home
So, my current directory is on a filesystem that's mounted on /dev/md0.
Based on the device that the filesystem is mounted on, you can then figure out if it's a local filesystem, an NFS mount, or something else.
Data deleted from /mnt directory after stop and start EC2 instance
[root#localhost opt]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/xvda1 9.7G 1.4G 7.8G 15% /
none 1.9G 0 1.9G 0% /dev/shm
/dev/xvdb 394G 199M 374G 1% /mnt
I place my data in /mnt .I stop instance yesterday .
Afetr starting the instance today ,I didnt find any data form /mnt .
I have another from /opt .
How can i recover that data from /mnt .
If /mnt is Temporary mounting point .Then how can i use these all space
On EC2, /mnt directory is mounted to ephemeral storage.
After reboot or instance stop/start, all data is lost.
Please refer to this post.
It is a common misconception that a reboot/restart will wipe the ephemeral storage - this isn't true.
You can try it yourself and see.
What will is a stop/start - that actually deprovisions your VM, and then moves it to another host machine - which will have wiped ephemeral drive(s) and then starts it up with your root ebs (at least) attached. Stop/start and reboot are often conflated - but they are very different things here.
/mnt should really just be used for ephemeral data storage that is not critical if instance needs to be restarted. This is actually well suited for things such as local on-disk cache, temporary data storage, etc. as this ephemeral storage will oftentimes perform better from an I/O standpoint than say an EBS volume mount. Just understand that you should only place non-critical data there.
I am relatively new to Linux. In one of our projects, we use amazon's EC2 instance for processing of some files. We upload files to S3 server after processing. EC2 instance is booted using an existing AMI
Recently I got an error no space left on disk, hence processing of files was halted. I cleaned up some older files and the processing continued.
Now when I look at available space using df -h
Filesystem Size Used Avail Use% Mounted on
/dev/xvda1 9.9G 5.7G 3.7G 61% /
none 3.7G 0 3.7G 0% /dev/shm
/dev/xvdb 414G 199M 393G 1% /mnt
/dev/xvdc 414G 199M 393G 1% /data
I can see my files are effecting only /dev/xvda1.
I have following queries
What is the use of other partitions when I can see my files only effecting /dev/xvda1
It looks like we are only using 10 GB of space effectively and other is being wasted. How can I use other space? Can I move some disk space to /dev/xvda1 or directly store files in other areas?
As you can see from the output of df -h, there are two large partitions mouted on /mnt and /data respectively. I suggest that you use those partitions by processing the files in one of those directories. If you cannot move where the processing happens for some reason, you can remount the partitions in the appropriate place.
If for example your files are processed in the directory /var/mydir and you cannot change that, do the following (as root):
umount /mnt
mount /dev/xvdb /var/mydir
You can use the other partition as well of course if you prefer that.