I currently have a 5-node triple mode FDB 7.1.27 cluster. I created a backup folder on one of the machines and changed the permissions before performing the backup.
mkdir /fdb/backup
chmod 777 /fdb/backup
fdbbackup start -d file:///fdb/backup/ -t local
The result after executing the above command is as follows:
backup]# fdbbackup start -d file:///fdb/backup/ -t local --logdir .
The backup on tag `local' was successfully submitted.
backup]# fdbbackup status -t local
The backup on tag `local' is in progress to file:///fdb/backup/backup-2023-11-30-14-19-56.363443.
BackupUID: 831b2776a5fbaac26e630b6d0bdbf843
BackupURL: file:///fdb/backup/backup-2023-11-30-14-19-56.363443
Snapshot interval is 864000 seconds. The initial snapshot is still running.
Details:
LogBytes written - 0
RangeBytes written - 455520173
Last complete log version and timestamp - 105206242641, 2023/11/30.14:17:42+0800
Last complete snapshot version and timestamp - N/A, N/A
Current Snapshot start version and timestamp - 105206320414, 2023/11/30.14:17:42+0800
Expected snapshot end version and timestamp - 105206320414, 2023/11/30.14:17:42+0800
Backup supposed to stop at next snapshot completion - Yes
Recent Errors (since initialization)
28 second(s) ago : 'File could not be written' on 'file_backup_write_range_5.2'
You can see that there is an error message “File could not be written’ on ‘file_backup_write_range_5.2’”. But the data directory has already been created.
backup]# ll
total 16
drwxr-xr-x 5 foundationdb foundationdb 67 Nov 30 14:27 backup-2023-11-30-14-19-56.363443
-rw-r--r-- 1 root root 14387 Nov 30 14:19 trace.100.71.8.128.43376.1701325196.U2Nc8k.0.1.xml
backup]# cd backup-2023-11-30-14-19-56.363443/
backup-2023-11-30-14-19-56.363443]# ll
total 0
drwxr-xr-x 3 foundationdb foundationdb 49 Nov 30 14:19 kvranges
drwxr-xr-x 3 foundationdb foundationdb 26 Nov 30 14:21 logs
drwxr-xr-x 2 foundationdb foundationdb 67 Nov 30 14:27 snapshots
backup-2023-11-30-14-19-56.363443]# du -sh ./*
2.0G ./kvranges
0 ./logs
20K ./snapshots
Does this have an impact?
In addition, if the same backup directory is created on any B node in the cluster and permissions are granted. Executing the backup command at the current node will also generate a backup file for node B.
nodeB backup]# ll
total 4
nodeB 4096 Nov 30 14:25 backup-2023-11-30-14-19-56.363443
nodeB backup]# ll backup-2023-11-30-14-19-56.363443
total 8
drwxr-xr-x. 3 foundationdb foundationdb 4096 Nov 30 14:25 kvranges
drwxr-xr-x. 3 foundationdb foundationdb 4096 Nov 30 14:25 logs
nodeB backup]# du -sh ./backup-2023-11-30-14-19-56.363443/*
195M ./backup-2023-11-30-14-19-56.363443/kvranges
12K ./backup-2023-11-30-14-19-56.363443/logs
nodeB backup]#
So do we need to create backup directories on all nodes in the cluster?
Thanks!