Preconditions: * clone the Ostro-OS project * build the swupd image and data (bitbake ostro-image-swupd) Expected outcome: The ETA reported is reasonably accurate Actual Outcome: The ETA reported is too optimistic. Example: At some point the progress was showing for about 15-20 minutes the status pasted below: Currently 2 running tasks (4403 of 4422) 99% |################################################################################################# | 0: bundle-ostro-image-swupd-mega-1.0-r0 do_rootfs (pid 35487) 81% |#################################################### | ETA: 0:00:25 1: bundle-ostro-image-swupd-world-dev-1.0-r0 do_rootfs (pid 35490) 81% |################################################ | ETA: 0:00:25
I found out that in some other cases the forecast is completely off by days: Initialising tasks: 100% |##########################################################################################################| Time: 0:00:07 NOTE: Executing SetScene Tasks NOTE: Executing RunQueue Tasks Currently 4 running tasks (4399 of 4422) 99% |################################################################################################# | 0: bundle-ostro-image-swupd-world-dev-1.0-r0 do_rootfs (pid 39704) 0% | | ETA: 11 days, 14:26:48 1: bundle-ostro-image-swupd-mega-1.0-r0 do_rootfs (pid 39701) 0% | | ETA: 11 days, 14:26:41 2: bundle-ostro-image-swupd-qa-bundle-b-1.0-r0 do_rootfs (pid 39705) 81% |############################################## | ETA: 0:00:36 3: bundle-ostro-image-swupd-qa-bundle-a-1.0-r0 do_rootfs (pid 39706) 81% |############################################## | ETA: 0:00:36 Notice the "ETA: 11 days, 14:26:48"
Based on this and another report I've had, I've just sent a patch to drop the ETA display for tasks altogether, since it's unlikely we'll be able to make it accurate all of the time with the information we have.
I think a fair compromise would be to show historical data (how long it took on average to perform task X) and elapsed time. The user can do own prediction.
We don't really have anywhere to get that historical data though - if we did then we'd have a reasonable way to actually produce an estimate. ETA removed in master: http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=5f4559b2eb8cbf633ee2c923c0b5e23a24b1ec06