# using xtrabackup on production servers

**URL:** <https://forums.percona.com/t/using-xtrabackup-on-production-servers/6668>\
**Category:** Percona XtraBackup\
**Created:** [November 6, 2018, 5:05am UTC](https://forums.percona.com/t/using-xtrabackup-on-production-servers/6668 "2018-11-06T05:05:28Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![unix196](https://avatars.discourse-cdn.com/v4/letter/u/0ea827/32.png) [@unix196](https://forums.percona.com/u/unix196)\
**Post date:** [November 6, 2018, 5:05am UTC](https://forums.percona.com/t/using-xtrabackup-on-production-servers/6668/1 "2018-11-06T05:05:28Z")

</div>

Good day! I have some question about using xtrabackup on loaded servers. Characteristics this server

```auto

debian stretch
percona-server-server-5.7 5.7.21-21-3.stretch
percona-xtrabackup-24 2.4.12-1.stretch
hardware memory 96G
16 core
load average - 10-12

mysql:
innodb_buffer_pool_size = 72G
459G - size of the db
QPS - 6500

```

We planning migrate backup database from mysqldump)) to xtrabackup =\> try test backup on production server.

```auto
[TABLE="border: 0, cellpadding: 0, cellspacing: 0, height: 1"]
[TR]
[TD="bgcolor: #f5f5f5"]
		[/TR]
[TR]
[TD="bgcolor: #f5f5f5"]
		[/TR]
[TR]
[TD="bgcolor: #f5f5f5"]
		[/TR]
[/TABLE]

```

datadir - is SSD, target-dir - HDD  
I try use ionice and throttle, but during backup our monitoring system + haproxy balancer peridically reported that database is unavailable (in nagios - checks fall of with timeout, haproxy - report that backend is unavailable) + ssh terminal on this database server peridically begin slow down.  
Try backup on other production database server - there the same problem.

I got the impression that xtrabackup must be using on slave server without user load (i.e. slave server for backup purpose ). Is it true?