Venus: [bug] Storage power stuck at 1 sector

[yangjian102621]

Well, i had sealed 13 GB data on the miner node, but it always displayed 268.4 MB on the https://stats.kittyhawk.wtf

miner@xjxh:~$ du -sh .filecoin*
0   .filecoin
6.5G    .filecoin-20190716-010345-v002
13G .filecoin_sectors

And the miner power is no changed after i had finished the first sector, it is always 268435456 / 2229088026624.

My miner node information

• ID : QmVoqEgQhLLSWMzLjaeofD4bAcpPNuakVaSw72YsK8Y7YZ
• minerAddress t2zmkmdccog3da2n5rnmkyrzvvuoklbcpymcapxay

I had post more than 60 deals to my miner with two client nodes, and my .filecoin_sectors dir is increasing 256 MB every 20 mins, but the data of https://stats.kittyhawk.wtf is never changed.

If someone could tell me what's problem?

[anorth]

268.4MB sounds suspiciously like exactly one sector. Despite sealing more sectors, I think that your PoSt is not being accepted. The PoSt is necessary to gain power.

This is very likely due to https://github.com/filecoin-project/rust-fil-proofs/issues/738. Let's use this as a tracking issue in go-filecoin for this problem.

[anorth]

There are reports from another miner about the same problem in slack.

[travisperson]

This actually has to do with the way the proving set is updated. We only update the proving set during a submitPoSt.

There is a special case for the first commitSector where we immediately add that sector to the proving set and start the proving period. This ensures miners get power quicker. This is the first submitPoSt that shows up on chain for any miner.

Due to how fast PoSt proof generation is at the moment you are unlikely to actually get another sector committed before the submitPoSt. This results in the next proving period to also only contain this first sector. This is why power does not change after the second submitPost. However, your proving set is probably much larger now, and on the third submitPoSt you should see your power increase.

Some details:

Your first commitSector was in block 13080, your first submitPoSt was in block 13093, and your second submitPoSt was in 14150

The proving period for your miner at the moment is 15080 to 16080. So we should see your power increase pretty significantly in a little while. The PoSt generation will take a little longer though, as we currently only do a PoSt over 2 sectors at a time.

[steven004]

Your first commitSector was in block 13080, your first submitPoSt was in block 13093, and your second submitPoSt was in 14150

@travisperson , I am a little confused by this. I just checked the submitPost message this particular miner sent in block 13093 and 14150 respectively, however, I saw the parameter length of the two messges are quite different: the first one is 268Bytes, while the second one is 2080. My understanding is that the length of params basically reflect the size of proving-set, so I would expect more proven storage will show up after the 2nd submitPost message successfully accepted by the chain. Is that correct?

[travisperson]

My understanding is that the length of params basically reflect the size of proving-set

That is correct. This is a bug at the moment which is being fixed in #3067. Currently when we start PoSt generation we take all committed sectors and submit all the proofs. However, the the protocol is only going to award power for the proving set sectors.

[yangjian102621]

@steven004

How to query the submitPost of specified miner?
Thanks.

[steven004]

@yangjian102621 just go to filecoin explorer.

[a8159236]

@yangjian102621 Hello, has your problem been solved? I had the same problem.I've already sent many data to the miners and shown "complete", but the size of "go-filecoin miner power" has not increased, and the "Storage Power" on the dashboard has not increased, for several days. What's the reason?Can you give me some guidance?

[yangjian102621]

@a8159236
Yep, i found my second submitPoSt at block 14148

This submitPoSt parameter length is 2080 Bytes, much more longer than the first one(268 Bytes),
my miner's storage power increase pretty significantly(increased to 16374562816 / 3143916060672).

Proven Storage is also increase to 16.4 GB at https://stats.kittyhawk.wtf/storage/mining

So you just to wait you miner node's second submitPoSt, then you "Storage Power" will increase a lot.

[a8159236]

@yangjian102621 How to query the submitPost of specified miner on filecoin explorer?Do you use the browser's own search function ，Or is it a one-by-one information query?

[yangjian102621]

@a8159236

How to query the submitPost of specified miner on filecoin explorer?Do you use the browser's own search function ，Or is it a one-by-one information query?

Sorry, i also have no idea, This question you ask @travisperson may be more appropriate.

[a8159236]

@yangjian102621
Well, thank you very much.

[yangjian102621]

@a8159236
Hi, If you have no obstacles to reading Chinese, here is a blog explaining how to check the submitPost of the specified miner, hope make a little help for you.

[a8159236]

@yangjian102621 Thank you very much. Can you provide this blog?

[yangjian102621]

@a8159236
Sorry, i forgot it. The blog url is http://www.r9it.com/20190720/filecoin-miner-test.html

[a8159236]

@yangjian102621 Thank you very much. But do you know what these two words mean respectively?createChannel and commitSector .

[a8159236]

@yangjian102621 The longer the length of the submitted parameters, the greater the additional storage power.But what factors determine the length of the parameters?
Are you Chinese? Is it convenient to leave the contact information of Wechat? I am also a fan of filecoin and hope to communicate more.

[a8159236]

Why did my parameter length reach 10KB?

[yangjian102621]

@a8159236
Sorry, this issue is closed.
If you have other questions to discuss, just send an email to my box : [email protected]

[yangjian102621]

@travisperson
Hi, is it possible to use the same minerAddress for two different nodes?

[anorth]

No, that is not going to work if the nodes are storing different data.

[taylorshuang]

No, that is not going to work if the nodes are storing different data.

@anorth what do you mean? I don't get it.

[taylorshuang]

This issue is very useful, thanks.

[yangjian102621]

@anorth
Thanks , i got it.
Another reason is that it is not possible to determine which node to post the data to sealing?

[yangjian102621]

@travisperson

Hi, I found there is an an API client for Filecoin implemented in javascript.
js-filecoin-api-client

Do you have plans to develop a filecoin api client for java implementation and PHP implementation?

If you do had a plan , i want to make a tittle contribute for it , i launched a project, an API client for Filecoin implemented in Java,(The project has just started, there are not many functions, but I intend to constantly improve it.)

we had a plan to develop a version implemented In PHP if you needed.

I want to add my project to filecoin-project, so that more people can use and participate in the contribution.
What can I do to achieve this?

Thanks.

[anorth]

@yangjian102621 this isn't the right place for such a discussion - I suggest the #fil-dev channel in the project slack, or the discussion forum.

[yangjian102621]

@anorth
I'm sorry, in fact i had created a top in discussion forum. But i did not got a reply,
and i had got a reply just now, we will discuss it on the forum.
Thanks for reply.