Temp dirs issue?

[bmentink]

HI,

Love the software, make's plotting much easier. However ..

I am getting the error: "no eligible tempdirs (5617s/5100s)" when I plot.

I am only running two parallel plots with a 1TB SSD. I would have thought that since temporary storage takes 280GB per plot, there would be enough temp space ..

Originally posted by @bmentink in #725

[altendky]

You will need to share your complete configuration file and the output from whatever mechanism you use to check the status of the plots. Plots take peak 250GB and usually more like 200 on average, or perhaps a bit less. I regularly run four on 1TB drives. But, plotman doesn't do that count for you. You have to configure it.

[byeuji]

Pasting what I wrote in the other thread:

If you're plotting in parallel or using the tmp2 feature, it might just be a temporary issue. What plotman probably means is "no eligible tempdirs right now"

You might also have the max jobs tmpdir_max_jobs setting too low -- with 1TB plotting drive, you should be able to set this to 4, since using plotman, not all jobs will be in Phase 2 at the same time. But personally, I'd recommend using the tmp2 feature of chia to avoid running out of space on tmp and gumming up your plots.

[bmentink]

@byeuji Thanks. I can't run 4 plots as I am memory limited (8GB). I would of thought that setting tmpdir_max_jobs to 2 would be fine when I am only ever running two plots at a time ..

@altendky As above , I am only ever using two plots at a time, so temp storage should not get above 500GB, so how is it I am getting the warning when I have 1TB, does not make sense ..

Relevant part of config is:

scheduling:
        # Run a job on a particular temp dir only if the number of existing jobs
        # before [tmpdir_stagger_phase_major : tmpdir_stagger_phase_minor]
        # is less than tmpdir_stagger_phase_limit.
        # Phase major corresponds to the plot phase, phase minor corresponds to
        # the table or table pair in sequence, phase limit corresponds to
        # the number of plots allowed before [phase major : phase minor].
        # e.g, with default settings, a new plot will start only when your plot
        # reaches phase [2 : 1] on your temp drive. This setting takes precidence
        # over global_stagger_m
        tmpdir_stagger_phase_major: 2
        tmpdir_stagger_phase_minor: 1
        # Optional: default is 1
        tmpdir_stagger_phase_limit: 1

        # Don't run more than this many jobs at a time on a single temp dir.
        tmpdir_max_jobs: 2

        # Don't run more than this many jobs at a time in total.
        global_max_jobs: 2

        # Don't run any jobs (across all temp dirs) more often than this, in minutes.
        global_stagger_m: 65

        # How often the daemon wakes to consider starting a new plot job, in seconds.
        polling_time_s: 20


plotting:
        k: 32
        e: False             # Use -e plotting option
        n_threads: 6         # Threads per job
        n_buckets: 128       # Number of buckets to split data into
        job_buffer: 3400     # Per job memory

EDIT: Is latest screenshot of warning ..

[altendky]

If you have one tmp dir set to allow two plots and it has two active plots running on it... then it is ineligible for starting a new plot. This sounds like just a misunderstanding of what the message means. Though, the message isn't present in your last screenshot. plotman doesn't pay attention to the total space or available space of a tmp dir.

[altendky]

(also, share the full config when requested, not the "relevant" part)

[bmentink]

Though, the message isn't present in your last screenshot.

Yes, it is .. 2nd text line from the bottom.

The message is misleading, and there is no clue as to what the numbers represent i.e (6263s/3900s) what is "s"?

[altendky]

s is seconds, that's the stagger. Also, that's a message from about 45 minutes before your screenshot. The active plotting launching status at the time of the screenshot is on the first line and is stagger (2505s/3900s) which means you are 2505 seconds through the configured 3900 second stagger period.

[bmentink]

Got it, thanks..

…

On Mon, 7 Jun 2021, 1:22 pm Kyle Altendorf, ***@***.***> wrote: s is seconds, that's the stagger. Also, that's a message from about 45 minutes before your screenshot. The active plotting launching status at the time of the screenshot is on the first line and is stagger (2505s/3900s) which means you are 2505 seconds through the configured 3900 second stagger period. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#730 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACNHVBZA2RT4ANODJJCZD5TTRQNOVANCNFSM46CYQPHA> .

[evynovev]

hello,
I have same message. process is running but is slow....is there any relation?

[altendky]

I don't think so particularly. If your plots are going slower than you expect then maybe try the #plotting channel on Chia's keybase. Or swing by #plotman there as well.