We’re using InfluxDB OSS v2.7.11 to replicate data from a local Raspberry Pi 5 (running Debian/Ubuntu ARM64) to a remote VM, using the built-in OSS-to-OSS replication feature.
This setup has worked reliably for 1–2 months, but for the last 3 days, replication is failing continuously, and queues are building up in each replication stream.
Problem
Replication queue has been increasing steadily. Replication status shows errors like:
read: connection reset by peer
context deadline exceeded (Client.Timeout exceeded while awaiting headers)
i/o timeout
curl http://<remote-ip>:8086/health = “pass”),
opsight@opsight:~/Documents/Press1 $ curl http://:8086/health
{“name”:“influxdb”, “message”:“ready for queries and writes”, “status”:“pass”, “checks”:, “version”: “v2.7.11”, “commit”: “fbf5d4ab5e”}
Replication Queue Status
Each replication shows increasing Current Queue Bytes and Remaining Bytes to be Synced, like:
opsight@opsight:~/Documents/Press1 $ influx replication list --org “NTF_UNIT5” -t “omCtv7hPJTdAqpyVEOvflKU06hcn7THMqKxCMC-bfIz4mmcPdja3V6oP6DRU_2EATP8fMq6vsCzj07XqvI1umg==”
ID Name Remaining Bytes to be Synced Current Queue Bytes on Disk Max Queue Bytes Latest Status Code
0ef447e3a0f5f000 abcShift_2_CSM3Shift_202c8f4e13d168612b2172153a05403ed 0 8 1073741824 204
0ef446a49bb5f000 abcToday_CSM3Today55405d9f658b340810cebe31f9f9a82e 47876141 48702680 1073741824 204
0ef447434675f000 abcShift_1_CSM3Shift_188a930951f61ea2d90cb47a30ef7cd06 11180083 15394350 1073741824 204
0ef447c08475f000 abcShift_2_STShift_295d93f419a668f21b2172153a05403ed 0 8 1073741824 204
0ef446c6afb5f000 abcToday_STTodayd5c8f3eed8bcedd210cebe31f9f9a82e 0 158237 1073741824 204
0ef447663a75f000 abcShift_1_STShift_1f25e69811c9cf35690cb47a30ef7cd06 0 158237 1073741824 204
Setup Details
| Item | Value |
|---|---|
| Source | Raspberry Pi 5 (ARM64) |
| OS | Ubuntu/Debian (64-bit) |
| InfluxDB | OSS 2.7.11 |
| Remote DB | InfluxDB OSS 2.7.11 |
| Data Rate | ~150 points per bucket every 30 seconds |
| Replication | Native OSS-to-OSS via influx replication create |
| Status | Broken for 3 days continuously |
Things Verified
-
Ping + HTTP access to remote InfluxDB
-
/healthon remote Influx returns"pass" -
Remote system has sufficient memory and disk
-
No recent config changes by us
-
Also, when I write data directly to the remote VM, it is successfully writing the data without any delay or issue, at the same time getting the same issue with the replication, regardless of whether the replication is old or new.
