Nodes are already running v25, but Mainnet is still v24—this time the discrepancy is different

12 days have passed since the V25 node deadline (6/18), and we've observed a new situation, different from the previous 'complete inactivity'.

Node container: v25       ✓ Running (pulled around 6/27–6/28)
Testnet 2 protocol: v24   ⏳ Not yet switched
Mainnet protocol: v24     ⏳ Not yet switched

The nodes themselves have automatically pulled and are running the v25 container. This is inferred from the container creation time found via docker images, which was approximately two to three days ago (around 6/27–6/28) – a 9–10 day discrepancy from the official 6/18 deadline. However, both Testnet 2 and Mainnet protocol_versions are currently still v24.

This Differs from Last Week's Observations

In our previous post, we noted 'complete inactivity.' Now, it appears that the node side is actually active: the v25 image has been pulled, and the container is running. However, this 'node software layer readiness' and the 'official on-chain protocol version switch' are two distinct stages – a pattern consistent with previous upgrades.

It's just that the gap between node-side readiness and the on-chain switch appears to be longer this time compared to previous instances.

Comparing Time Lags with the Past

V22  Node Deadline 4/27 → Mainnet Switch 4/30 (3-day difference)
V23  Node Deadline 5/19 → Mainnet Switch 5/22 (3-day difference)
V24  Node Deadline 6/2  → Mainnet Switch 6/5 (3-day difference)
V25  Node Deadline 6/18 → Mainnet Switch ? (12 days passed, no switch)

The past three switches were completed within 3 days. This time, the node side is ready (v25 container started running around 6/27–6/28), but as of this publication, 12 days have passed since the deadline, and no on-chain change has been observed – this discrepancy is significantly longer than the previous three.

We are unsure of the reason. It could be that there are more conditions to consider for this switch, or it might simply not be time yet – we have no official documentation to confirm the specific reason and will not invent explanations here.

Also Noted: New Pi Node Software Version 0.6.0

This is separate from the container pull mentioned above, but the timing coincides within the last few days, so we're noting it together.

Version 0.6.0 of the Pi Node software is currently available for download on the official node page. This update is not automatically pushed and requires manual download and installation from the official website. We have not yet upgraded to this version, so we are currently unaware of the specific features added or adjusted in 0.6.0.

This is not causally related to the automatic v25 container pull mentioned earlier – we are currently running an older version of the Pi Node software, and the v25 container was automatically pulled and is running on this older version. In other words, you don't need to upgrade to 0.6.0 to get v25. What 0.6.0 brings remains to be investigated.

Current Knowns and Unknowns

Knowns:

Node software (older version) has automatically pulled and is running the v25 container, around the deadline.
Testnet 2 and Mainnet protocol_versions are still v24.
A new version of Pi Node software, 0.6.0, is available for manual download; its content awaits verification.

Unknowns:

Why the on-chain switch is slower than the previous three times.
What specific features 0.6.0 adds.

OrbitWatch Monitoring Continues

Our Horizon API monitoring continues to track the Mainnet protocol_version. We will publish an update once the on-chain version switches or when specific update details for 0.6.0 are found.

Further Reading

Data sources: Direct observation of node docker images, Pi Network Horizon API (api.mainnet.minepi.com), Pi Network Block Explorer, minepi.com/pi-node official website. All analyses do not constitute investment advice. OrbitWatch is an independent Pi Network ecosystem observatory and is not affiliated with the official Pi Network.