Thanks to everybody who submitted bug reports and feature requests for ec2-consistent-snapshot, software which can be used to create consistent EBS snapshots on Amazon EC2 especially for use with XFS and/or MySQL.
A new release (0.1-9) has been published to the Alestic PPA. This release provides the following fixes and enhancements:
Read MySQL “host” parameter from .my.cnf if provided. (closes: bug#485978)
Support quoted values in .my.cnf (closes: bug#454184)
Require Net::Amazon::EC2 version 0.11 or higher. (closes: bug#490686)
Require xfsprogs package. (closes: bug#493420)
Replace “/etc/init.d/mysql stop” with “mysqladmin shutdown”. (closes: bug#497557)
Document —description option which was added earlier. (closes: bug#487692)
If you already have ec2-consistent-snapshot installed, you can upgrade using commands like:
sudo apt-get update
sudo apt-get install ec2-consistent-snapshot
If you don’t yet have the Alestic PPA set up, run these commands first:
codename=$(lsb_release -cs)
echo "deb http://ppa.launchpad.net/alestic/ppa/ubuntu $codename main"|
sudo tee /etc/apt/sources.list.d/alestic-ppa.list
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys BE09C571
If you find bugs or think of feature requests, please submit them.


Eric,
Thanks for all the great work with this utility. I'm using it in several production systems. It's been a real time saver
-David
Thanks for creating this as well as everything else you've created. Very helpful for developers trying to get started with serious EC2 setups. I've tried to create my first snapshot using this script. I have a mysql database running on a raid of 4 volumes. This database is slaving from another. When I run ec2-consistent-snapshot, I get the following:
ec2-consistent-snapshot \
> --region us-east-1 \
> --xfs-filesystem /mnt/db \
> --mysql \
> vol-1 vol-2 vol-3 vol-4
ec2-consistent-snapshot: master_log_file="mysql-bin.000467", master_log_pos=98
ec2-consistent-snapshot: ERROR: create_snapshot: Attribute (progress) does not pass the type constraint because: Validation failed for 'Str' failed with value undef at /usr/local/share/perl/5.10.0/Net/Amazon/EC2.pm line 828
Net::Amazon::EC2::create_snapshot(undef, 'VolumeId', 'vol-1', 'Description', 'ec2-consistent-snapshot') called at /usr/bin/ec2-consistent-snapshot line 136
eval {...} called at /usr/bin/ec2-consistent-snapshot line 134
main::ebs_snapshot('ARRAY(0xa34dc50)', 'us-east-1', 'ec2-consistent-snapshot') called at /usr/bin/ec2-consistent-snapshot line 90
Yes, I changed the volume ids above. The first snapshot gets created, but everything else dies. Any thoughts?
myles: This is a known defect in Net::Amazon::EC2 version 0.12. Jeff Kim is releasing a new version which will correct this problem, so keep an eye on CPAN and upgrade when it's available.
http://search.cpan.org/~jkim/Net-Amazon-EC2/
Hello,
Thanks for creating this tool. I am planning to use it to backup my ubuntu EC2 instance.
I have been getting errors with and without using sudo. With sudo, it throws up error about keys and takes no snapshot. Without sudo, it throws up error about xfs_freeze, but does creates a snapshot.
Any help will be much appreciated.
regards,
anil
ubuntu@domU-12-31-38-00-0D-41:/mnt2/aks$ sudo ec2-consistent-snapshot --xfs-filesystem /mnt2 --description "mnt2 source" vol-f3ba039a
ec2-consistent-snapshot: ERROR: Can't find AWS access key or secret access key at /usr/bin/ec2-consistent-snapshot line 76.
xfs_freeze: cannot unfreeze filesystem mounted at /mnt2: Invalid argument
ec2-consistent-snapshot: ERROR: xfs_freeze -u /mnt2: failed(256)
ubuntu@domU-12-31-38-00-0D-41:/mnt2/aks$ ec2-consistent-snapshot --xfs-filesystem /mnt2 --description "mnt2 source" vol-f3ba039a
xfs_freeze: cannot freeze filesystem at /mnt2: Operation not permitted
ec2-consistent-snapshot: ERROR: xfs_freeze -f /mnt2: failed(256)
snap-7c432e14
xfs_freeze: cannot unfreeze filesystem mounted at /mnt2: Operation not permitted
ec2-consistent-snapshot: ERROR: xfs_freeze -u /mnt2: failed(256)
Anil:
By default sudo will clear your environment which may include the pointers to your AWS credentials.
Try adding the -E option to pass environment variables in to the command:
Alternatively, you could put your AWS credentials in one of the files that ec2-consistent-snapshot looks for.
Much gratitude for all your work Eric - you're a titan openness and developer education.
Question:
I have a 10.04 EBS boot AMI with a single XFS-formatted root volume containing a LAMP stack. I want to snap the entire root file system (including MySQL) for backup purposes, and it's working using:
sudo ec2-consistent-snapshot --aws-access-key-id XXXXXXXXXXXXXX --aws-secret-access-key XXXXXXXXXXXXXX --xfs-filesystem / --description "root_snapshot $(date +'%Y-%m-%d %H:%M:%S')" vol-XXXXXXX
Do I gain any benefit/consistency by including the --mysql option, or the mysql username and password options? Or is it unnecessary because I'm snapping the entire root file system?
Thanks again.
Ben
Ben: The --mysql option causes ec2-consistent-snapshot to flush the database changes to disk before initiating the snapshot. This helps ensure that the database files are consistent.
We have been using a precursor of this technique in a production system for about 18 months. Suddenly MySQL started "going away" with clockwork every time the script ran. MySQL support concluded "the 'SYSTEM ec2-create-snapshot $VOLUME_NAME' did not return in a hurry / got suck and hence filesystem access stayed suspended."
A reboot of the server was necessary to recover.
Is this something others have experienced? Is this potential lockup solved with ec2-consistent-snapshot?
The exact script we were running is:
#!/bin/sh
export EC2_HOME=/root/ec2-api-tools-1.3-24159
export EC2_PRIVATE_KEY=/root/.ec2/pk-*********.pem
export EC2_CERT=/root/.ec2/cert-********.pem
export JAVA_HOME=/usr/lib/jvm/java-6-openjdk
export PATH=$PATH:$EC2_HOME/bin
MYSQLUSER=UUUUUUU
MYSQLPASS=PPPPPPP
# Obtain instance data
INSTANCE_NAME=`curl -s http://169.254.169.254/2008-02-01/meta-data/instance-id`
echo "Instance: $INSTANCE_NAME"
VOLUME_NAME=`ec2-describe-volumes | grep $INSTANCE_NAME | grep -o "vol-\S*"`
echo "Volume: $VOLUME_NAME"
VOLUME_LOCATION=/mnt/data
mysql -u$MYSQLUSER -p$MYSQLPASS FLUSH TABLES WITH READ LOCK;
SHOW MASTER STATUS;
SYSTEM xfs_freeze -f /mnt/data
SYSTEM ec2-create-snapshot $VOLUME_NAME
SYSTEM xfs_freeze -u /mnt/data
UNLOCK TABLES;
EXIT
EOF
# Delete older snapshots (oldest 2).
NUMBER_SNAPSHOTS=`ec2-describe-snapshots | grep $VOLUME_NAME | wc -l` echo "Snapshots: $NUMBER_SNAPSHOTS"
if [ "$NUMBER_SNAPSHOTS" -gt "200" ]; then
# Delete oldest two.
ec2-describe-snapshots | grep $VOLUME_NAME | sort -k 5 | head -2 | awk '{print "Deleting: " $2; system("ec2-delete-snapshot " $2)}'
#ec2-describe-snapshots | grep $VOLUME_NAME | sort -k 5 | head -2 | awk '{print "Deleting: " $2}'
fi
Eric:
ec2-consistent-snapshot may help avoid the locking problem. The script has a lot of timeout and retry operations to prevent issues like this.