I searched " plots refresh" its not in my yaml.
I changed " plot loading frequency seconds"
I use gui, so never use init.
Thatās the old and not used anymore value that is left in config for no reason (on updates). You need to add that new section:
plots_refresh_parameter:
interval_seconds: 120 # The interval in seconds to refresh the plot file manager
retry_invalid_seconds: 1200 # How long to wait before re-trying plots which failed to load
batch_size: 300 # How many plot files the harvester processes before it waits batch_sleep_milliseconds
batch_sleep_milliseconds: 1 # Milliseconds the harvester sleeps between batch processing
If I update will it be there? donāt really wanna be adding bits to my yaml.
If you want to change that scanning duration, you have to add that to your config.
As long, as Chia will not change it again, it will stay there, regardless of new updates. The problem is with ignoring the original value and not inserting new values on updates. As @Fuzeguy stated, when you remove your config, it may add it there, as it starts from scratch. However, that may be again a different code path than a clean install, so may have other issues. Unfortunately, manual is the only safe way in this case (IMO).
By the way, new plots / drives that are added to harvester are discovered, based on those values. Therefore, when you increase it a bit, and you want to see a change immediately, you will need to refresh plots in UI.
1 Like
Let me know how it goes, I dont wanna mess with things unless I know its gonna help.
1 Like
A nice comparison between Chia GUI and Flexfarmer
Look at the difference between left and right. Chia gui makes an excessive amount of reads/writes to your disc that causes quite a few issues.
Could you name those issues?
Main one would likely be delayed lookups.
You mean that reading from a local disk takes over 5 seconds to check for one partial?
My understanding is that most people have sub 1 sec checks. Some checks are slow (e.g., problem being chased in this thread), but I would not imply that those are due to āmore complex plot structure,ā thus slower reads that your farmer can address.
On the other hand, some people persist to use network attached drives, and suffer from constant problems. Some people have misconfigured setups. As mentioned before, you would help more by better addressing those people while on their famers page.
It is hard for me to understand, why for instance on Stats tab, those āStaleā and āInvalidā are not in red. Why those fields donāt have tooltips like āReportedā to guide people. Why when you refresh page, you donāt play a sound (optional) when stales/invalid are detected.
I guess, we are off topic here, and I am starting to rant again.
Are these 2 charts really comparable?
1 is looking at i/o per second, the other is looking at amount of data read and written.
So no of actions vs size of actions.
A chart actually showing i/o for both, or amount of data written / read by both would be comparable, but thats not what Iām seeing here.
24 hours in it switches on flexfarmer on each chart.
1 Like
I am now two days without issue. So something Iāve done has fixed it for me.
1 Like
I have to hope this is not chias solution to our problems, seems a very poor fix.
Extract from 1.2.8 change log.
- A change to logging to only log warnings when more than 10 seconds has passed, to reduce the number of warning logs.
Not sure it even worked as planned considering how many are now getting spendbundle warnings for 1 sec.
That might be the result of this addition, but Iām not sureā¦from the changelogā¦
- Starting logging the timing of applying additions and removals to the coin store.
And esp. as Iām still getting less than 10 sec warnings about missing āLooking up qualitiesā, I have doubts that it even actually changed anything.
- change to logging to only log warnings when more than 10 seconds has passed, to reduce the number of warning logs.
1 Like
Well, Iām gonna call all my analysis worthwhile.
I certainly had driver issues.
Iām still getting warnings, but nowt in event viewer.
My warnings are greatly reduced, and the few Iām getting are only @ just over the 5 sec mark, which shouldnāt deny me a reward.
Maybe I hit the limit on my old pcās.
Curious how many plots you peeps have if you donāt mind sharing that info.
Iāve just topped 500 TiB.
Would love to add more, so curious if your also working with a similar amount.
I could run a spare machine as a harvester, but they take alot of power vs adding more drives to my set up.
I have some piās in my spares wardrobe, but if an 8350 struggles Iām not confident theyāll be a great help.
Did one change, and since 10am (so more than 10 hrs) this morning, not a single WARNING about disk not responding in time. Update tomorrow, hoping 
1 Like
@Bones Thanks so much for sharing your thoughts and possible solutions. seeing this thread has brought some sanity back to my insane long nights Iāve been having trying to resolve the same issues.
I should have come here sooner to look for others with the same problem.
@lowestofttim I see you suffer as well, but resolved it, hope I can do the same.
I did not have any issues, up until updating my GUI to 1.2.8 , did not have any warnings before that, not even one. And now my log is full of them. So im pretty confident its from the new gui but had no luck solving it.
Will try your suggestions and come back in 24+ hours to report my āluckā at this game of pushing different buttons at random 
1 Like
Well, got thru over 12hr - no lookup warnings, then sadly 12am, 1am, 3am, 5am it started again. What I had ādoneā new, was to upgrade to the latest Beta sw of GUI. Oh well 
Iām leaning towards, us mere mortals canāt fix this.
I have 1 drive that needs a defrag , so Iāll buy another and move the plots to sort that, but Iām kinda outta ideas.
Maybe Iāll also try using 2 machines vs the 1.
What are your times like, just over 5 sec or worse?