Side Load Restore Failure

,
1

I have a series of daily backups that I would like to load into on InfluxDB instance.

I used Influx Version 1.7.4 to archive the data.

My restore node uses Version 1.7.6

I can restore the first day without issue, since the database does not exist in my restore node.

$ influxd restore -portable -db telegraf /tmp/influx-archives/2019-05-19
2019/06/05 17:46:21 Restoring shard 63 live from backup 20190520T010302Z.s63.tar.gz
2019/06/05 17:46:21 Restoring shard 14 live from backup 20190520T010302Z.s14.tar.gz
2019/06/05 17:46:21 Restoring shard 54 live from backup 20190520T010302Z.s54.tar.gz
2019/06/05 17:46:21 Restoring shard 81 live from backup 20190520T010302Z.s81.tar.gz
2019/06/05 17:46:21 Restoring shard 26 live from backup 20190520T010302Z.s26.tar.gz
2019/06/05 17:46:21 Restoring shard 35 live from backup 20190520T010302Z.s35.tar.gz
2019/06/05 17:46:21 Restoring shard 2 live from backup 20190520T010302Z.s2.tar.gz
2019/06/05 17:46:21 Restoring shard 72 live from backup 20190520T010302Z.s72.tar.gz
2019/06/05 17:46:23 Restoring shard 6 live from backup 20190520T010302Z.s6.tar.gz
2019/06/05 17:46:23 Restoring shard 45 live from backup 20190520T010302Z.s45.tar.gz

The database exists and has data.

$ influx
Connected to http://localhost:8086 version 1.7.6
InfluxDB shell version: 1.7.6
Enter an InfluxQL query
> USE telegraf
Using database telegraf
> SELECT count(usage_guest) FROM cpu 
name: cpu
time count
---- -----
0    157860
>

I then restore the next day of data into telegraf_bak:

$ influxd restore -portable -db telegraf -newdb telegraf_bak /tmp/influx-archives/2019-05-20
2019/06/05 17:51:10 Restoring shard 35 live from backup 20190521T010302Z.s35.tar.gz
2019/06/05 17:51:10 Restoring shard 2 live from backup 20190521T010302Z.s2.tar.gz
2019/06/05 17:51:10 Restoring shard 26 live from backup 20190521T010302Z.s26.tar.gz
2019/06/05 17:51:10 Restoring shard 63 live from backup 20190521T010302Z.s63.tar.gz
2019/06/05 17:51:10 Restoring shard 6 live from backup 20190521T010302Z.s6.tar.gz
2019/06/05 17:51:10 Restoring shard 14 live from backup 20190521T010302Z.s14.tar.gz
2019/06/05 17:51:10 Restoring shard 45 live from backup 20190521T010302Z.s45.tar.gz
2019/06/05 17:51:10 Restoring shard 81 live from backup 20190521T010302Z.s81.tar.gz
2019/06/05 17:51:12 Restoring shard 54 live from backup 20190521T010302Z.s54.tar.gz
2019/06/05 17:51:12 Restoring shard 72 live from backup 20190521T010302Z.s72.tar.gz

The new telegraf_bak exists and includes data.

$ influx
Connected to http://localhost:8086 version 1.7.6
InfluxDB shell version: 1.7.6
Enter an InfluxQL query
> USE telegraf_bak
Using database telegraf_bak
> SELECT count(usage_guest) FROM cpu 
name: cpu
time count
---- -----
0    159138
>

I then copy over the data following the documented sideload procedure.

> [centos@ip-172-31-90-245 ~]$ influx
Connected to http://localhost:8086 version 1.7.6
InfluxDB shell version: 1.7.6
Enter an InfluxQL query
> USE telegraf_bak
Using database telegraf_bak
> SELECT * INTO telegraf..:MEASUREMENT FROM /.*/ GROUP BY *
ERR: no data received
>

Any idea why I get ERR: no data received?

2

I used CloudFormation to deploy an EC2 instance for my restore node.

CloudFormation sets a default disk size of 8GB.

I redeployed with a 150GB drive and the command worked.

> USE telegraf_bak
Using database telegraf_bak
> SELECT count(usage_guest) FROM cpu
name: cpu
time count
---- -----
0    157860
> SELECT * INTO telegraf..:MEASUREMENT FROM /.*/ GROUP BY *
ERR: partial write: field type conflict: input field "usage_system" on measurement "docker_container_cpu" is type float, already exists as type integer dropped=6800
> USE telegraf
Using database telegraf
> SELECT count(usage_guest) FROM cpu
name: cpu
time count
---- -----
0    315504
>