influxdb: Backup failure: Download shard n failed copy backup to file

Steps to reproduce: Execute a backup command like this: influxd backup -database mydb -portable -since 2018-01-01T00:00:00Z backup_$(date +%Y%m%d_%H%M%S)

Expected behavior: Get a backup written to files.

Actual behavior: Fails to generate a backup. Console output is as follows.

2020/02/05 14:40:47 backing up db=mydb
2020/02/05 14:40:47 backing up db=mydb rp=my_rp shard=122 to backup_20200205_144047/mydb.my_rp.00122.00 since 2018-01-01T00:00:00Z
2020/02/05 14:40:47 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (0)...
2020/02/05 14:40:49 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (1)...
2020/02/05 14:40:52 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (2)...
2020/02/05 14:40:54 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (3)...
2020/02/05 14:40:56 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (4)...
2020/02/05 14:40:58 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2s and retrying (5)...
2020/02/05 14:41:01 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 3.01s and retrying (6)...
2020/02/05 14:41:04 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 11.441s and retrying (7)...
2020/02/05 14:41:15 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 43.477s and retrying (8)...
2020/02/05 14:41:59 Download shard 122 failed copy backup to file: err=<nil>, n=0.  Waiting 2m45.216s and retrying (9)...
2020/02/05 14:44:44 error (copy backup to file: err=<nil>, n=0) when backing up db: mydb, rp my_rp, shard 122. continuing backup on remaining shards
2020/02/05 14:44:44 backup failed: copy backup to file: err=<nil>, n=0
backup: copy backup to file: err=<nil>, n=0 

Only once have I gotten a backup, and that was when the database was newly-created.

• The directory that contains the shard noted in the output above contained hundreds of empty temp directories. Deleting those directories didn’t change anything with the backup failure.
• Of the 132 shards, one has 8 tsm files, the problem shard has one tsm file and the others have one tsm file.
• Running influx_inspect commands verify and inspect show no issues.

Environment info:

• System info: Linux 4.15.0-1066-azure x86_64
• InfluxDB version: InfluxDB v1.7.0 (git: 1.7 dac4c6f571662c63dc0d73346787b8c7f113222a)
• Other relevant environment details: Influx instance is running on a pod in a AKS cluster. Database files are on a mounted Azure File share.

Config: Non-default configs are reporting-disabled = true bind-address = "0.0.0.0:8088" auth-enabled = true