Details

    • Type: Umbrella Umbrella
    • Status: Open
    • Priority: Major Major
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: documentation, regionserver
    • Labels:
      None

      Description

      We have been working on the ability to do backups in HBase with different levels of protection. This is an umbrella task for all the backup related changes. Here are some kinds of changes - will create separate issues for them:

      Roughly here are a few flavors of backups giving increasing levels of guarentees:
      1. Per cf backups
      2. Multi-cf backups with row atomicity preserved
      3. Multi-cf backups with row atomicity and point in time recovery.

      On the perf dimension, here is a list of improvements:
      1. Copy the files - regular hadoop "cp"
      2. Use fast copy - copy blocks and stitch them together, saves top of rack bandwidth
      3. Use fast copy with hard links - no file copy, it does only ext3 level linking.

      On the durability of data side:
      1. Ability to backup data onto the same racks as those running HBase
      2. Intra-datacenter backup
      3. Inter datacenter backup

      Restores:
      1. Restore with a table name different from the backed up table name
      2. Restore a backed up table wen HBase cluster is not running at restore time
      3. Restore into a live and running cluster

      Operationally:
      1. How to setup backups in live cluster
      2. Setting up intra-DC
      3. cross-DC backups
      4. Verifying a backup is good

        Issue Links

          Activity

          Karthik Ranganathan created issue -
          Karthik Ranganathan made changes -
          Field Original Value New Value
          Description We have been working on the ability to do backups in HBase with different levels of protection. This is an umbrella task for all the backup related changes.

          Roughly here are a few flavors of backups giving increasing levels of guarentees:

          1. Per cf backups
          2. Multi-cf backups with row atomicity preserved
          3. Multi-cf backups with row atomicity and point in time recovery.

          On the perf dimension, here is a list of improvements:

          1. Copy the files - regular hadoop "cp"
          2. Use fast copy - copy blocks and stitch them together, saves top of rack bandwidth
          3. Use fast copy with hard links - no file copy, it does only ext3 level linking.
          We have been working on the ability to do backups in HBase with different levels of protection. This is an umbrella task for all the backup related changes. Here are some kinds of changes - will create separate issues for them:

          Roughly here are a few flavors of backups giving increasing levels of guarentees:
          1. Per cf backups
          2. Multi-cf backups with row atomicity preserved
          3. Multi-cf backups with row atomicity and point in time recovery.

          On the perf dimension, here is a list of improvements:
          1. Copy the files - regular hadoop "cp"
          2. Use fast copy - copy blocks and stitch them together, saves top of rack bandwidth
          3. Use fast copy with hard links - no file copy, it does only ext3 level linking.

          On the durability of data side:
          1. Ability to backup data onto the same racks as those running HBase
          2. Intra-datacenter backup
          3. Inter datacenter backup

          Restores:
          1. Restore with a table name different from the backed up table name
          2. Restore a backed up table wen HBase cluster is not running at restore time
          3. Restore into a live and running cluster

          Operationally:
          1. How to setup backups in live cluster
          2. Setting up intra-DC
          3. cross-DC backups
          4. Verifying a backup is good
          Karthik Ranganathan made changes -
          Link This issue requires HBASE-4591 [ HBASE-4591 ]
          Nicolas Spiegelberg made changes -
          Link This issue blocks HBASE-4907 [ HBASE-4907 ]

            People

            • Assignee:
              Karthik Ranganathan
              Reporter:
              Karthik Ranganathan
            • Votes:
              0 Vote for this issue
              Watchers:
              32 Start watching this issue

              Dates

              • Created:
                Updated:

                Development