Kusama AssetHub - Collator appears inactive in on-chain storage and Polkadot.js UI but is actively producing blocks

[CoinStudioDOT]

Hi,

I’ve been a long-time collator on Kusama’s Asset Hub, and over the last few days I’ve noticed some strange behaviour.
I’m not sure whether it’s related to the recent migration or not, but it looks like there’s a mismatch between on-chain state and live network activity.

---

Based on the Polkadot.js Apps active collator list, my node COINSTUDIO/SYSTEM is not listed as active:
Active Collators (Polkadot.js)

However, if you observe the live network activity, you can clearly see that I am active and signing blocks:
Explorer View (Polkadot.js)

This suggests that it’s not just a Polkadot.js display issue, but something related to on-chain events or storage.

---

The first time I noticed this, I was unsure of the cause and tried executing a
collatorSelection.leave_intent extrinsic, followed by re-registering as a candidate.

However, the leave intent extrinsic failed, which seems consistent with a faulty on-chain state — my collator was marked as inactive in storage, while still actively authoring blocks.

After re-registering, I reappeared in the collator list with an expected block height for activation. On my first assigned block, a
CollatorSelection(CandidateRemoved) event
was triggered, and on the next assignment, my node resumed normal block production.

This can also be confirmed here:
🔗 Subscan Account Activity

---

It looks like:

• Checking on-chain storage does not show my collator as active,
• Yet querying recent blocks returns blocks authored by my collator,

---

Could someone please investigate whether there’s a discrepancy between the collatorSelection storage, on-chain event logic, and actual authoring behaviour after the recent migration?

[skunert]

This suggests that it’s not just a Polkadot.js display issue, but something related to on-chain events or storage.

The active list on pjs does not show you collator, but your account CdW8izFcLeicL3zZUQaC3a39AGeNSTgc9Jb5E5sjREPryA2 is indeed shown in the list of validators for the session where you produced block 11221188.

[
 CbLdvHtKPVr1bvt7poBKwJsQUokfjC4kqoMkjeyCDtFs8wQ
 E5G5qLNBF7nwhemokqNrNwUCQ7oHrLsBVjSc7XhTdRq8672
 EPGHRDFaBbQz4uF8DgnuScxYp4rvvtpL5MTed7VvGprNQmm
 EPk1wv1TvVFfsiG73YLuLAtGacfPmojyJKvmifobBzUTxFv
 Eicei1RVHvqHBPJAY2fZpLy2KtCMjY7Kwjo3svjmGLH7FmT
 Et9M3rrA7H2kHQEGRXHxufcp9HTEmFirMWtKHvjoJ85r1C9
 H1tAQMm3eizGcmpAhL9aA9gR844kZpQfkU7pkmMiLx9jSzE
 HPUEzi4v3YJmhBfSbcGEFFiNKPAGVnGkfDiUzBNTR7j1CxT
 HWyLYmpW68JGJYoVJcot6JQ1CJbtUQeTdxfY1kUTsvGCB1r
 J11Rp4mjz3vRb2DL51HqRGRjhuEQRyXgtuFskebXb8zMZ9s
 JL21EURyqQxJk9inVW7iuexJNzzuV7HpZJVxQrY8BzwFiTJ
 HmatizNhXrZtXwQK2LfntvjCy3x1EuKs1WnRQ6CP3KkNfmA
 DtuntvQBh9vajFTnd42aTTCiuCyY3ep6EVwhhPji2ejyyhW
 DV7b5AL9zNbsnQ8CWCwvw2tzc3pz74EVDLGQdNU1tujA2L5
 === Here
 CdW8izFcLeicL3zZUQaC3a39AGeNSTgc9Jb5E5sjREPryA2
 FPJeeKkKhDE8uyid4MKvTj7KJNVczkobiDsxuMmrjPA4Wec
 EXTRGokoMvv1BKrRKhjsz97tjHo8RCkNc1xcEQukLBN2NNH
]

So maybe its indeed a display bug in the list of active collators shown on pjs.

[seadanda]

Just had a quick check (in transit) - subscan shows you have been removed from the candidate list a few times on init. It could be that you're not making blocks in time and get removed, or that there's some other issue with your collator. Bond seems fine on first glance and it doesn't appear that anybody is calling take_slot on you, and your leave intents failed as you were already out of the list.

[CoinStudioDOT]

Just had a quick check (in transit) - subscan shows you have been removed from the candidate list a few times on init. It could be that you're not making blocks in time and get removed, or that there's some other issue with your collator. Bond seems fine on first glance and it doesn't appear that anybody is calling take_slot on you, and your leave intents failed as you were already out of the list.

It's strange that after the candidate removed on init I continue to sign blocks. This morning I actually dropped and now issued take_candidate slot extrinsic. Will see what will happen.

[skunert]

Just had a quick check (in transit) - subscan shows you have been removed from the candidate list a few times on init. It could be that you're not making blocks in time and get removed, or that there's some other issue with your collator. Bond seems fine on first glance and it doesn't appear that anybody is calling take_slot on you, and your leave intents failed as you were already out of the list.

It's strange that after the candidate removed on init I continue to sign blocks. This morning I actually dropped and now issued take_candidate slot extrinsic. Will see what will happen.

Hmm but I mean the CandidateRemoved event will be emitted immediately, but the real list of authorities will only be updated when the SessionManager is invoked next. So should only come into effect at the next session?

[CoinStudioDOT]

Thanks for clarifying that.

In any case, the sequence of events still seems a bit unusual, particularly the one where the CandidateRemoved event was emitted right on the first block assignment, after which I started collating normally again.

The exact cause is unclear, but it looks like several things happened simultaneously, possibly including other collator registering for the slot (with much lower bond), assuming it was empty since my node wasnt shown in the active list .

After executing the take candidate slot extrinsic, everything appears to have returned to normal which was not the case when I tried the regular path of registering as a candidate and then updating the bond to have a competitive amount.

I am not sure there is much more to extract from this right now, maybe we can close the issue for now, and if the same behaviour occurs again, we can revisit it with more data.

Thanks for looking into this.