Describe the bug

Hi All.
I attempted to upgrade my 2.19.2 cluster to 3.3.0. I completed this on two other small clusters without issue. However, on this one the first data node failed to recover

java.lang.IllegalArgumentException: cannot change field "@timestamp" from docValuesSkipIndexType=NONE to inconsistent docValuesSkipIndexType=RANGE

{
"index": ".ds-xxxxxx-000038",
"shard": 1,
"primary": false,
"current_state": "unassigned",
"unassigned_info": {
"reason": "ALLOCATION_FAILED",
"at": "2025-10-16T18:33:30.400Z",
"failed_allocation_attempts": 5,
"details": """failed shard on node [bPDnq6B3RvKmX9DHyEl8DQ]: shard failure, reason [index id[qxZL7pkBSC1xwYofggg_] origin[PEER_RECOVERY] seq#[6425736]], failure IllegalArgumentException[cannot change field "@timestamp" from docValuesSkipIndexType=NONE to inconsistent docValuesSkipIndexType=RANGE]""",
"last_allocation_status": "no_attempt"
},

i have removed the data from the disk and re-deployed 2.19.2 container, allowing it to replicate data back onto it from other nodes - this works as expected.

can someone help identify why these shards are failing - i have many. They are all data streams using @timestamp. Mapping for timestamps:

  "properties": {
    "@timestamp": {
      "type": "date"
    },

Related component

No response

To Reproduce

1. difficult as i have upgraded 2x clusters from 2.19.2 to 3.2.0 without issue.
2. @timestamp is mapped to date in these index's which from the log above looks to be the cause for some reason. this works on 2.19.2

Expected behavior

expect shards to run in 3.3.0 as they do in 2.19.2. Cluster was initially built on 2.11 and upgraded to most 2.x releases to 2.19 without issue.

Additional Details

Plugins
Please list all plugins currently enabled.

Screenshots
If applicable, add screenshots to help explain your problem.

Host/Environment (please complete the following information):

• OS: [e.g. iOS]
• Version [e.g. 22]

Additional context
Add any other context about the problem here.