Friday, November 15, 2019

something i was able to isolate this afternoon is that the laptop only runs at 30 w if it's left on, or at least is only running at 30 w now that i've disabled everything i can. there does seem to be a small spike when it turns on, but it only comes out to around 0.01 or 0.02 added, so it's kind of just error. i was thinking it would be more than that...

i'll get a better bounds on the modem tomorrow.

so, if i leave it on all of the time, as i have been, i'm using

0.03(12*.105 + 6*.205 + 6*.144) = $0.10/day.

if, on the other hand, i'm strict about turning it off from 7:00-19:00, i'd be using:

0.03*12*.105 + .02 = $0.05-$0.06/day, which is just half the total amount.

if i leave it on for peak period and turn it off at night, it's:

0.03*6*(.205 + .144) + 0.02 =  $0.08/day

if i average that out, the difference is $0.02-$0.03 cents a day, or $0.60-$.90/month. so, let's say about $0.75/month.

it's not worth being neurotic about, i don't think.

i should just turn it off when i'm not using it. if that averages six hours a day, it's 5-6 kwh
https://wikileaks.org/vault7/document/SeaPea-User_Guide/SeaPea-User_Guide.pdf
today was completely lost, but in the end may not be wasted.

i was midway through writing the following post at around 9:15 this morning when i made a terrible error that resulted in me losing the day but may fix the access issues i was having on the drive.

this is what i was writing...

so, i was very, very slow getting started today.

i checked my email minutes after the post at 3:26 this morning and spent the next two hours dealing with that, including applying for a second mastercard through pc. they're threatening to shut down the site, now, if they don't get something in three days. i feel like i'm under an extortion attack by islamic terrorists - "give us the money in three days or it's all over". fuck.

i had to write several emails trying to order the thing out, and all i can do is hope it works.

but, it means i need to call td today. i'm in for the weekend, but will probably do an outside run on monday. that might include stopping by the bank and having a chat.

i used to have all of this information in boxes, but it's one of the things that disappeared when i left my items in storage. i have virtually everything from 2011-on, still. the credit karma site says the last payment was in june, 2006. really. i'm asking about a 13 year-old credit card. well, it's what google wants, what can i say...

i'll have to call the board today, too.

i then spent some more time thinking about the fan. can i just buy a 20 w desk fan? nobody seems to want that. 45 w? maybe if i import it from europe.

to finish the thought, i then experimented a little to see if i can plug the usb fan right into the wall, and i can, but i'm not getting as much amperage from it.

the next thing to check was my usage from the other day, and i noticed a spike where i didn't expect one (in hindsight, i can explain the spike). i thought i was sleeping (i wasn't). did i turn my laptop back on (yes)? so, i went to check the event log and couldn't open the file.

at that point, i realized that the entire system32 folder was owned by an alphanumeric s-12367tg77567896yp90 user. so, i went to take ownership and did, and then tried to remove the user, but couldn't. it said i had to remove the inheritance. i didn't think it through, i just clicked ok.

all of my security tabs disappeared.

fuck.

well, i can try a reboot, right? nope...

so, i was back in winpe on the desktop, and while the chkdsk was quick, the startup recovery couldn't find the operating system.

thankfully, i had a good understanding of what i did and the background to fix it (i did vista support, remember). just something else i'm technically an expert in. so, if you ever find yourself in the situation where you remove all ownership to a directory, accidentally or not, what you need to do is run the following script:

for /l %%x in (1, 1, n) do (
   echo %%x
   takeown /f dir /r /a /d y
   icacls dir /inheritance:e /t /q
   icacls dir /reset /t /q  
)

...where n is some appropriate number of iterations, dir is the directory.

if you do this for a system directory, as i did, n must be large, and you get something like this:

for /l %%x in (1, 1, 200) do (
   echo %%x
   takeown /f d:\windows\system32 /r /a /d y
   icacls d:\windows\system32 /inheritance:e /t /q
   icacls d:\windows\system32 /reset /t /q  
)

 so, i let that run and took a nap.

and, it still wasn't done when i woke up.

what i'm doing with the script is (1) recursively taking ownership of every file in the directory tree and putting it in the admin group, (2) turning inheritance back on and (3) resetting the acl to default. (1) will error out on each iteration, until it's done. (2) and (3) will then increase the size of the numbers processed over the next iteration of (1), so the loop is necessary to complete the process, and, while this should end, it can get frustrating when you only get through a few files per iteration, once you've been through a few thousand, already, because you have to do them again.

so, i did the last few thousand files manually by jumping directories. and, it wasn't done until 17:00...

that was brutal. it's my fault. i wanted to crawl into a hole somewhere. but, i merely lost a few hours on something that i've seen stump mcses. this is considered abstract in the field of tech support. i think they think there's a better answer, but there just isn't, you have to use brute force. fwiw, while i didn't write the tests (because they were charging a lot for them, and i wanted to get paid when i went to work), i went through three months of training by a certified microsoft employee that they flew in to ottawa from seattle, and then worked second tier microsoft support for a few months (until they shipped the jobs to the philipines). he hated the weather and couldn't get out fast enough. i bet it's nicer in manilla, if you can deal with the hurricanes. but, i basically have an mcse on top of the other degrees. i usually forget to mention that when enumerating my unofficial degrees, but that's in there, too. so, i figured this out quickly because i know what i did and how to fix it. most tech support agents would have had a hard time with that, though - that's a higher tier support problem because file systems are considered to be an abstract concept.

anyways.

i logged back in and noticed that the funny username, which i had up to this point assumed was myself from the previous install, had disappeared, at least, but that doing so had left a stream of previously hidden files, almost as though i had a rootkit. upon closer inspection, these files appeared to be exclusively related to remote access.

i've been pointing out for months that i think the cops have something installed on the drive....

so, i spent a few hours doing things like removing services, setting group policy settings and deleting directories. in the end, i can't say that i have any evidence of intrusion, but it's funny - i've turned on the event log (which was off, and i don't remember turning off) as well as the group policy editor (same.), so i should now be able to track any intrusion, when i previously couldn't. to be clear: if the cops installed something on the drive, they also turned off the software that would log them doing it. if they try it again, i should catch them in the act.

and, that's the day.

i didn't call the board. i didn't call td. i didn't finish the liner notes from inri000.

but, i'm going to get something to eat and try to push myself to finish it before i fall asleep.
it's just too cold to go out this weekend. this is normal for me - i shut down when it gets cold, and come back in the spring. i may come out of hibernation to forage if it seems nice, but i usually see my shadow. there are exceptions for artists i can't miss - i saw do make say think on a cold night in december, 2017 - but they are few and far between.

like, there's some chance i could be out once or twice between now and april.

so, there's some shows on saturday, but i'm staying in. and, we'll see if the weather co-operates a bit better next week or not.

there's a very high chance i'll get to plaid. that's the only thing coming up right now.
i ended up sleeping until after 22:00. so, it wasn't merely a nap. i'm up, now, and have eaten, but i need to take a shower before i get back to finishing this - and i really should finish it tonight.

i did get my fan.

the device is about the size i expected, but it's substantively weaker than the one running, which is not exactly a surprise. i knew it would be weaker, but how much weaker? i cannot seem to get rid of the burning smell, which currently smells like a burning wood log, like at a camp fire, which is still gross, but at least a little better. the type of burning shifts, but the burning is constant. the fan is absolutely necessary. so, this is not going to replace the fan, or at least won't for the moment, and i'm going to need to find other ways to cut down electrical, including turning the modem off when it isn't in use.

i tested it on the two usb 3 ports on the chromebook and on the usb 2 ports on the laptop and the chromebook. i could not a discern a difference in pull. but, i got a lot less out of the hub (which would have wired the fan, the mouse, a usb key and the wireless keyboard in series - presumably on 0.1 A.).

i suppose that it's possible that the different ports are running in series, locally. that would mean i could have gotten up to 0.5A on the laptop, and again on the chromebook. but, if i could pull at a higher amperage, why wouldn't i have pulled it out of the usb 3 ports? the fact that i didn't would indicate i'm pulling at 0.5A on the device, max.

but, when plugged directly into the laptop, the fan doesn't seem to respond at all to changes in power consumption out of the other ports. it reacted to the keyboard in the hub, but not to the hard drive in the port next to it. i tried copying files to and from the drive, running media off of it - nothing affected this stoic fan, in the slightest. it just churns. keep in mind that the maximum power usage over usb 2.0 is only 2.5W, and an external drive is definitely going to use all of that if it can. that would suggest that the ports are really not in series at all, that it's getting it's maximum pull out of a 0.1 A connection and that it just can't pull any more than that. ever.

so, the device would be 0.5W max, rather than 2-3. that's kind of what i thought. tech specs are terrible across the board. you don't know what you're buying, anymore. if elizabeth warren wants to run against technology, why doesn't she insist on regulating specs so that you actually get what you pay for?

regardless, this is enough evidence to take a gamble on the solar panel, if i think it can do well in low light. i actually think that reducing the circuitry could increase the amount i can pull, if there's some kind of breaker being pushed down by the os. it might actually pull more power from the leds, if i can get a solid 250-300 mA signal. and, will that turn it up a little?

logic is fun, and logic is useful, but logic is not enough for a proper epistemology - i do defer to empiricism when required. i want a reading before i plan my next move, here. so, i got one of those little ammeter/voltmeter combos for what was left on the pre-paid debit ($11). i overpaid, but i don't want to wait a month for it to get here from china. it should be here by monday.

if i do the reading, and it's only pulling 100 mA then i've got the green light to go solar. if it's pulling higher than that after all, and this is the best the thing can do with sufficient power, then it's not worth it, after all, and i'm stuck.