Add troubleshooting guidance for 'shard group has no shards' write failures

Nobody has claimed this yet.

## Add troubleshooting guidance for "shard group has no shards" write failures

### Affected pages

**InfluxDB v2:**
- https://docs.influxdata.com/influxdb/v2/write-data/troubleshoot/
- https://docs.influxdata.com/influxdb/v2/reference/internals/shards/
- https://docs.influxdata.com/influxdb/v2/admin/backup-restore/restore/

**InfluxDB v1 OSS:**
- https://docs.influxdata.com/influxdb/v1/administration/backup_and_restore/
- https://docs.influxdata.com/influxdb/v1/guides/write_data/

**InfluxDB v1 Enterprise:**
- https://docs.influxdata.com/enterprise_influxdb/v1/administration/backup-and-restore/

### Change

**Write troubleshooting (v2)** — add a new section:

Currently, there is no documentation for the error `shard group N covering to has no shards`. Add a troubleshooting entry covering:

- **Error message**: `failure writing points to database: shard group N covering UTC to UTC has no shards`
- **Cause**: A shard group exists in the metadata (BoltDB) but contains zero shards. Reports are associated with changes in write precision or backup/restore operations.
- **Affected versions**: In versions prior to 1.12 and 2.7.12, this condition causes a panic (divide by zero). In 1.12+ and 2.7.12+, the server returns an error instead of panicking.
- **Behavior**: Writes to the affected time range fail until the shard group duration expires and a new shard group is created for the next time window. Writes to other time ranges are unaffected.
- **Workaround**: Wait for the affected shard group's time range to pass, or copy data from the affected bucket into a new bucket.
- **Tracking**: [GitHub issue #25715](https://github.com/influxdata/influxdb/issues/25715)

**Shards internals page (v2)** — add a note under "Shard groups":

Add a brief note that a shard group with zero shards is an abnormal state that causes write failures for that time range. Link to the troubleshooting entry.

**Restore pages (v2, v1, enterprise)** — add a caution:

Under the restore instructions, add a note that in rare cases, backup/restore operations may produce shard groups with zero shards, which causes write failures. Link to the troubleshooting entry (or describe inline for v1 pages).

**v1 write_data page** — add a troubleshooting note:

Document the error message and link to the relevant GitHub issue for v1 users encountering this after backup/restore.

### Constraints
- Applies to InfluxDB v1 (1.x) and InfluxDB v2 (2.x) — both use the TSM storage engine with shard groups
- Does not apply to InfluxDB 3 (different storage engine, no shard groups)
- Do not state a definitive root cause. Use language like "reports are associated with" rather than "caused by"
- Internal metadata repair tools exist but are not shipped and should not be documented

### Source analysis
- PR [#25717](https://github.com/influxdata/influxdb/pull/25717) (merged Dec 2024) prevents the divide-by-zero panic when a shard group has no shards. Ported to 2.7 branch via [#26389](https://github.com/influxdata/influxdb/pull/26389) and to 1.x via [#25718](https://github.com/influxdata/influxdb/issues/25718).
- Issue [#25715](https://github.com/influxdata/influxdb/issues/25715) tracks the root cause investigation.
- PR [#27035](https://github.com/influxdata/influxdb/pull/27035) (open) adds logging to detect null-shard groups during backup/restore.
- Engineering confirmed: fix is in 1.12+ and 2.7.12+. Writes fail until shard group duration expires.

---
Source: DAR issue (internal), source code analysis

Review the listed InfluxDB v2, v1 OSS, and Enterprise troubleshooting, shard, write, and restore pages first. Add the requested troubleshooting entry, shard-group note, and restore cautions without asserting a definitive root cause; link the relevant tracking issue where specified. Done means all affected pages cover the zero-shard write failure, versions, behavior, workaround, and InfluxDB 3 exclusion.

Written by the indexing model from the issue text.