# Percona Cluster Backups - Process for recovery

**URL:** <https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617>\
**Category:** Percona XtraBackup\
**Created:** [May 11, 2020, 3:55pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617 "2020-05-11T15:55:06Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![meyerder](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/meyerder/32/16994_2.png) [@meyerder](https://forums.percona.com/u/meyerder)\
**Post date:** [May 11, 2020, 3:55pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617/1 "2020-05-11T15:55:06Z")

</div>

I am attempting to figure out what I did wrong on my dev system  
I wanted to try out a recovery from backups on a cluster and appear to have failed…&nbsp; Not quite sure why… gcache.size=4G on a pretty slow transaction server…&nbsp; should not have exceeded this valuewsrep\_desync=off  
mkdir /backupinnobackupex --galera-info backup  
innobackupex --apply-log .  
  
rm -rf /clusterdatadir  
manually created grastate.dat from xtrabackup\_galera\_info  
# GALERA saved stateversion: 2.1uuid:&nbsp; &nbsp; 15b2b8d2-46a8-11ea-a8de-8e1112674a6d:18413827seqno:&nbsp; &nbsp;23cert\_index:  
cp -R . /clusterdatadir/cd /clusterdatadirchown -R mysql:mysql .service mysql start  
This started a entire sync from another node in the cluster vs picking up where it should have…

---

<div class="post-metadata">

**Author:** ![meyerder](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/meyerder/32/16994_2.png) [@meyerder](https://forums.percona.com/u/meyerder)\
**Post date:** [May 11, 2020, 4:21pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617/2 "2020-05-11T16:21:59Z")

</div>

I may have found it… I missed in the uid that the sequence number is also included…

---

<div class="post-metadata">

**Author:** ![meyerder](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/meyerder/32/16994_2.png) [@meyerder](https://forums.percona.com/u/meyerder)\
**Post date:** [May 11, 2020, 4:36pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617/3 "2020-05-11T16:36:28Z")

</div>

Yep that was my issue 🙂

---

<div class="post-metadata">

**Author:** ![meyerder](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/meyerder/32/16994_2.png) [@meyerder](https://forums.percona.com/u/meyerder)\
**Post date:** [May 11, 2020, 4:53pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617/4 "2020-05-11T16:53:03Z")

</div>

The only remaining question is let us assume I lost the ENTIRE cluster… how to get it to start as a single node like it is the “master”&nbsp;  
Do I nuke the grastate file… modify wsrep.conf to show only a single server set to bootstrap and bring up?

---

<div class="post-metadata">

**Author:** ![vadimtk](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/vadimtk/32/9528_2.png) [@vadimtk](https://forums.percona.com/u/vadimtk)\
**Post date:** [May 13, 2020, 3:40pm UTC](https://forums.percona.com/t/percona-cluster-backups-process-for-recovery/7617/5 "2020-05-13T15:40:32Z")

</div>

Hi, your plan looks correct.  
Anyway - if you have a backup you can try it out. If it does not work - we will help you to figure this out.