# Copyright (c) 2006-2009 Mitch Garnaat http://garnaat.org/ # # Permission is hereby granted, free of charge, to any person obtaining a # copy of this software and associated documentation files (the # "Software"), to deal in the Software without restriction, including # without limitation the rights to use, copy, modify, merge, publish, dis- # tribute, sublicense, and/or sell copies of the Software, and to permit # persons to whom the Software is furnished to do so, subject to the fol- # lowing conditions: # # The above copyright notice and this permission notice shall be included # in all copies or substantial portions of the Software. # # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS # OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABIL- # ITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT # SHALL THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, # WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS # IN THE SOFTWARE. # """ Automated installer to attach, format and mount an EBS volume. This installer assumes that you want the volume formatted as an XFS file system. To drive this installer, you need the following section in the boto config passed to the new instance. You also need to install dateutil by listing python-dateutil in the list of packages to be installed in the Pyami seciont of your boto config file. If there is already a device mounted at the specified mount point, the installer assumes that it is the ephemeral drive and unmounts it, remounts it as /tmp and chmods it to 777. Config file section:: [EBS] volume_id = logical_volume_name = device = mount_point = """ import boto from boto.manage.volume import Volume from boto.exception import EC2ResponseError import os, time from boto.pyami.installers.ubuntu.installer import Installer from string import Template BackupScriptTemplate = """#!/usr/bin/env python # Backup EBS volume import boto from boto.pyami.scriptbase import ScriptBase import traceback class Backup(ScriptBase): def main(self): try: ec2 = boto.connect_ec2() self.run("/usr/sbin/xfs_freeze -f ${mount_point}", exit_on_error = True) snapshot = ec2.create_snapshot('${volume_id}') boto.log.info("Snapshot created: %s " % snapshot) except Exception as e: self.notify(subject="${instance_id} Backup Failed", body=traceback.format_exc()) boto.log.info("Snapshot created: ${volume_id}") except Exception as e: self.notify(subject="${instance_id} Backup Failed", body=traceback.format_exc()) finally: self.run("/usr/sbin/xfs_freeze -u ${mount_point}") if __name__ == "__main__": b = Backup() b.main() """ BackupCleanupScript= """#!/usr/bin/env python import boto from boto.manage.volume import Volume # Cleans Backups of EBS volumes for v in Volume.all(): v.trim_snapshots(True) """ TagBasedBackupCleanupScript= """#!/usr/bin/env python import boto # Cleans Backups of EBS volumes ec2 = boto.connect_ec2() ec2.trim_snapshots() """ class EBSInstaller(Installer): """ Set up the EBS stuff """ def __init__(self, config_file=None): super(EBSInstaller, self).__init__(config_file) self.instance_id = boto.config.get('Instance', 'instance-id') self.device = boto.config.get('EBS', 'device', '/dev/sdp') self.volume_id = boto.config.get('EBS', 'volume_id') self.logical_volume_name = boto.config.get('EBS', 'logical_volume_name') self.mount_point = boto.config.get('EBS', 'mount_point', '/ebs') def attach(self): ec2 = boto.connect_ec2() if self.logical_volume_name: # if a logical volume was specified, override the specified volume_id # (if there was one) with the current AWS volume for the logical volume: logical_volume = next(Volume.find(name=self.logical_volume_name)) self.volume_id = logical_volume._volume_id volume = ec2.get_all_volumes([self.volume_id])[0] # wait for the volume to be available. The volume may still be being created # from a snapshot. while volume.update() != 'available': boto.log.info('Volume %s not yet available. Current status = %s.' % (volume.id, volume.status)) time.sleep(5) instance = ec2.get_only_instances([self.instance_id])[0] attempt_attach = True while attempt_attach: try: ec2.attach_volume(self.volume_id, self.instance_id, self.device) attempt_attach = False except EC2ResponseError as e: if e.error_code != 'IncorrectState': # if there's an EC2ResonseError with the code set to IncorrectState, delay a bit for ec2 # to realize the instance is running, then try again. Otherwise, raise the error: boto.log.info('Attempt to attach the EBS volume %s to this instance (%s) returned %s. Trying again in a bit.' % (self.volume_id, self.instance_id, e.errors)) time.sleep(2) else: raise e boto.log.info('Attached volume %s to instance %s as device %s' % (self.volume_id, self.instance_id, self.device)) # now wait for the volume device to appear while not os.path.exists(self.device): boto.log.info('%s still does not exist, waiting 2 seconds' % self.device) time.sleep(2) def make_fs(self): boto.log.info('make_fs...') has_fs = self.run('fsck %s' % self.device) if has_fs != 0: self.run('mkfs -t xfs %s' % self.device) def create_backup_script(self): t = Template(BackupScriptTemplate) s = t.substitute(volume_id=self.volume_id, instance_id=self.instance_id, mount_point=self.mount_point) fp = open('/usr/local/bin/ebs_backup', 'w') fp.write(s) fp.close() self.run('chmod +x /usr/local/bin/ebs_backup') def create_backup_cleanup_script(self, use_tag_based_cleanup=False): fp = open('/usr/local/bin/ebs_backup_cleanup', 'w') if use_tag_based_cleanup: fp.write(TagBasedBackupCleanupScript) else: fp.write(BackupCleanupScript) fp.close() self.run('chmod +x /usr/local/bin/ebs_backup_cleanup') def handle_mount_point(self): boto.log.info('handle_mount_point') if not os.path.isdir(self.mount_point): boto.log.info('making directory') # mount directory doesn't exist so create it self.run("mkdir %s" % self.mount_point) else: boto.log.info('directory exists already') self.run('mount -l') lines = self.last_command.output.split('\n') for line in lines: t = line.split() if t and t[2] == self.mount_point: # something is already mounted at the mount point # unmount that and mount it as /tmp if t[0] != self.device: self.run('umount %s' % self.mount_point) self.run('mount %s /tmp' % t[0]) break self.run('chmod 777 /tmp') # Mount up our new EBS volume onto mount_point self.run("mount %s %s" % (self.device, self.mount_point)) self.run('xfs_growfs %s' % self.mount_point) def update_fstab(self): f = open("/etc/fstab", "a") f.write('%s\t%s\txfs\tdefaults 0 0\n' % (self.device, self.mount_point)) f.close() def install(self): # First, find and attach the volume self.attach() # Install the xfs tools self.run('apt-get -y install xfsprogs xfsdump') # Check to see if the filesystem was created or not self.make_fs() # create the /ebs directory for mounting self.handle_mount_point() # create the backup script self.create_backup_script() # Set up the backup script minute = boto.config.get('EBS', 'backup_cron_minute', '0') hour = boto.config.get('EBS', 'backup_cron_hour', '4,16') self.add_cron("ebs_backup", "/usr/local/bin/ebs_backup", minute=minute, hour=hour) # Set up the backup cleanup script minute = boto.config.get('EBS', 'backup_cleanup_cron_minute') hour = boto.config.get('EBS', 'backup_cleanup_cron_hour') if (minute is not None) and (hour is not None): # Snapshot clean up can either be done via the manage module, or via the new tag based # snapshot code, if the snapshots have been tagged with the name of the associated # volume. Check for the presence of the new configuration flag, and use the appropriate # cleanup method / script: use_tag_based_cleanup = boto.config.has_option('EBS', 'use_tag_based_snapshot_cleanup') self.create_backup_cleanup_script(use_tag_based_cleanup) self.add_cron("ebs_backup_cleanup", "/usr/local/bin/ebs_backup_cleanup", minute=minute, hour=hour) # Set up the fstab self.update_fstab() def main(self): if not os.path.exists(self.device): self.install() else: boto.log.info("Device %s is already attached, skipping EBS Installer" % self.device)