- takes 1.1% longer (2 min during a 3 hour backup)
- uses 1.4% more space (1G more in a 70G backup — this was for uncompressed backups)
- uses 1.115% more cpu overall
- split as 0.12% user, 0.66% nice, 0.025% system, 0.31% more iowait, 0% more steal
InnoDB Hot Backup:
- slightly fewer resources (see xtrabackup)
The server has 32G of RAM, 8 Dual-Core AMD Opteron(tm) Processor 8220, with a speed of 2800 Mhz (2.734 Ghz) and a cache size of 1024 Kb, and is also used as a reporting server. We ran Xtrabackup on odd days, and InnoDB hot backup on even days. Both are stable and have responsive developers, and both projects have developers that know the internals of MySQL.
The clients of ours that use InnoDB Hot Backup (including the client this test was run on) already have purchased indefinite licenses for it. For those clients, there is not a big reason to switch to Xtrabackup.
For companies that are not using a hot backup solution yet, we recommend both InnoDB Hot Backup and Xtrabackup, and they end up choosing what best fits their environment. Some companies shun InnoDB Hot Backup because of its price tag. Some companies are wary of Xtrabackup because it has been less than a year since the release was announced, and feel better paying money to a company for their backup solution. I did not put the “feelings” in the lists above, because they are dependent on the environment — I have the same trust in the skill and commitment of the developers of Xtrabackup and the developers of InnoDB Hot Backup. Other such feelings might be “supporting open source” or “using the backup software that is developed by a partner organization.”
We felt it was easier to set up InnoDB Hot Backup than Xtrabackup, but that may be because we have been using InnoDB Hot Backup for years. However, neither setup is too complex — it is not necessary to have a senior-level DBA set up the backups, no matter which software you choose.
17 Responses to “Taste test: Innobackup vs. Xtrabackup”
Leave a Reply