Is my farm broken or am I just bad luck?

I have the same “problem” here in Germany. I have about 4300 Plots, waiting time 14 days and since the update from 1.2.11 to 1.3.X at the beginning of march NO wins. Evething seem fine with my farm. I tried farmr looks good, i am faming in a pool now for 2 weeks looks good. I also synched from scratch. The times are below 1 sec, no errors in the logs…
In the past i used to win much often with half of plots that i have now and bigger net space. My biggest hit was 3 wins in 9 days, waiting time 3 month: D Pretty odd stuff.

While I’m waiting gives me more time to eat more of my favorite things.

1 Like

Just luck, i have apx 7000, started pooling as i went over 2 mnths with no blocks, apx time to win was 9 days.
Last mnth i hit 7 blocks so a lucky streak.
Im now at 24 days since i hit a block.

It all evens out in the end.

ETW of 2-3 days and it’s been almost 20 … I would be worried too. I can confirm that wins are possible with 1.3.5. I’m running it and won on Thursday. I’m not sure how you made it so long (such a large farm) with no monitoring software. I use Chiadog. It lets me know right away when there is an issue. You have a huge investment there. Take the time to setup a system to make sure it is running at peak performance all the time and that you know the moment something is wrong.

Good luck.

2 Likes

It looks like my sub slots is crap. You know how I can improve this?

Unfortunately i cant.

There is a post here from the farmr maker explaining it a little bit.

Can you post a pic of your farmr readout?

On mine i have 125 complete sub plots, what is yours at?

If you didnt have your log level set to info before installing farmr, you should expect your no to be lower until its been running set to info for 24 hrs.

1 Like

just limited down the peers to 30 to see if it will improve anything and this is from the last 10 in

Hmmm.
Your longest response time is far too long.
I get a tiny % over 5s and im ok with that, but 56 sec is far too long.
Missed 9 challenges also not good especially if your solo and not pooling.

I know restarts can throw those numbers, so hopefully thats why theyre looking so bad.

Hope the node lowering works for you.

FYI next to your name on farmr there is a down pointing arrow cliicking that and choosing legacy can show a slighy more in depth view.
No need to post it here, just wanted to let you know.

Agreed, that’s way out of bounds and should be investigated. The resulting missed challenges are a definite no-no, there is some real problem in the node.

However,

farmr ave

shows the min, avg, and median times are fine. They could (maybe should) be about 1/2 that (because it’s certainly possible), but 1.0s average is nothing to be ashamed of. Bottom line, the node doesn’t miss challenges too much, so, statistically, rewards shouldn’t be effected terrible much… but they are. Ummm.

To start troubleshooting, I would suggest opening your log file in notepad (or something) and looking for that time and see (possibly what HD is causing it) and find the others too, to see if there is a commonality.

Thanks guys.

that’s from reloading 24K plots. :rofl:

I do that get once or twice a day, it seems to be at a fix time of the day.

1 Like

As fuzeguy said, keep an eye on it , the log will tell you which plots are taking a long time, if there is a particular drive involved every time it could be dying or need looking at.

What would be the factor that affect “complete sub slot” and to improve this.
I did some google and it say 64 signage is 1 sub slot but how to make sure this is improve and best optimized?

Did you read the link i posted from gill?
It seems sub plots are not that important necessarily.

Once your node has been up for 24hrs without you messing with it, repost another screen shot from farmr pls.

Because if after 24 hrs without you messing with it you are still seeing really long responses or missed challenges i think that will be what needs investigating.

I have just under 1/3 of your plots. I run an old cpu ( fx 8350 ) and my times very very rarely go over 10 sec.

What are the specs of your node?
What is your setup, 1 main node im guessing?
How many harvesters?

Thank you for the help.
I will send another screenshot tomorrow.

I’m running 1 full node.
i7 9700 3.8Ghz
16Gb memory
Chia is on 500Gb SATA SSD
OS; Windows
connected to 4x JBOD via 2 x12Gb SAS card.
Each JBOD has a direct SAS link (not chained)

This machine is purely for Chia farming so nothing else is running on it.

I plot on other rig and move the HDD to the JBOD once full.
These rigs are not connected to the internet.

Nothing has changed on the machine but the last drop I got was 2nd May.
Just thought its been a while and I’m paranoid something is wrong.

thanks again

1 Like

I think that 1sec average may not be really the issue. If his distribution would be a perfect bell curve, that would imply that half of his proofs are below 1 sec, and half between 1 and 2 secs. Also, the median is 0.9 sec, that further suggest not that many are over 5 secs. We know that the right side is not perfect, but, it has to be close (with just few outliers over 5 secs).

My thinking was that moving plots over the wire was causing those bursts, but as HDs are moved, that is ruled out. This leaves us with either drives going to sleep, or having bad sectors. (Not sure what else could be.) If drives would be going to sleep, that would imply that drives are rather small (main part of the filter is done in memory), as such we should see a bunch of 10-20 secs there. What kind of suggests that maybe there are bad clusters on those drives.

Another option (don’t have a clue about it), is that those JBODs are acting up from time to time (some internal scan kicks in every so often???). Although, that may be not the case, as those are server grade devices, so performance is the king with those.

So, maybe chkdsk would be warranted to go over those drives. I (personally) would go for smartctl long scan (smartctl -t long /dev/sdb), as this one doesn’t really care about the file structure, but surface integrity. (chkdsk would be much faster, so I would start with it)

With such CPU, you should not need to reduce the number of peers, as that mostly is needed for slower CPUs and db on HDs. However, I would check in Task Manager your RAM utilization, and would get another 16GB of RAM.

By the way, I am on v1.3.3 and am also 3-4x behind schedule, and all looks clean. However, seeing this thread, I upgraded to v1.3.5 today. The initial v1.3.0 was messing up payout addresses, so maybe there were some code quirks with lower than v1.3.5.

Being 6-7x behind is not really being paranoid, as that should not happen too often. You have a big farm, so you should see more outliers, still that is kind of high.

1 Like

My high response would only pop up 2-3 times a day.
It seem to be on similar time of the day as well and they are all over random drives.

My last drop was on 1.3.3, 2nd May.
I was paranoid the older version is affecting my win so I updated to 1.3.4 on the 11th May. which the GUI was a mess, and lucky they release 1.3.5 on the 12th.

Same shit here since 1.3.4 / 1.3.5. Won last block like 6-8 weeks ago. Should win in 2-3 Weeks.

1 Like

The long win wait could be bad luck however…

This is an issue, even if a small one and IMO needs further investigation.

Im not saying its not to do with the software, but that is totally normal and will happen at times.

Exactly what I’ve seen, I won in May 2021, again in June, then started pooling late August as no wins and won a few days later on 23 August, then nothing until March this year (most that 6 months was 2 Month ETW), then two weeks later won again, and nothing since. I’ve had an ETW of 1 month since since the end of March.

I know my system is working well as I’m pooling and running Farmr, just the way it goes sometimes, but always worth checking everything through.

My brother runs two systems, one at his house the other at his girlfriends, not a 50/50 split of plotted space, but every win he’s had has been at his house, and I think he’s had five or six.

1 Like