| lotuspsychje | good morning to all | 04:02 |
|---|---|---|
| lordievader | Good morning | 07:04 |
| lotuspsychje | hey lordievader | 07:04 |
| lordievader | Hey lotuspsychje | 07:05 |
| lordievader | How are you doing? | 07:05 |
| lotuspsychje | great here, -4 again this morning | 07:05 |
| lordievader | Yeah, appearantly the weekend will be brutal. | 07:05 |
| lotuspsychje | lordievader: hggdh said they have -18 in texas | 07:06 |
| lotuspsychje | weather app says -7 monday | 07:07 |
| lordievader | Brr | 07:09 |
| ducasse | good morning, everyone | 07:41 |
| lotuspsychje | hey ducasse | 07:41 |
| ducasse | hi lotuspsychje - how's life today? | 07:42 |
| lotuspsychje | great here mate | 07:42 |
| lotuspsychje | still snowing there ducasse | 07:42 |
| ducasse | no, not any more, thankfully | 07:43 |
| lordievader | Hey ducasse | 08:47 |
| lordievader | How are you doing? | 08:47 |
| ducasse | hi lordievader - i'm fine, thanks, planning a trip out to have dinner and see a movie tonight. how are you? | 08:58 |
| lordievader | Sounds nice, what movie? | 08:58 |
| lordievader | I'm doing allright. | 08:58 |
| ducasse | a norwegian one, a comedy. it's a remake of a classic norwegian comedy, so might be fun. | 09:00 |
| jink | How are we ever going to find it? | 09:01 |
| jink | "A movie, about something, might be fun." | 09:01 |
| ducasse | http://www.imdb.com/title/tt7597486/ | 09:02 |
| lordievader | ducasse: Have fun 😉 | 09:14 |
| jink | :D | 09:14 |
| EriC^^ | morning all | 09:17 |
| jink | Morning, EriC^^. | 09:18 |
| ducasse | hi EriC^^ | 09:20 |
| akik | not sure if serious: https://bugs.launchpad.net/ubuntu/+source/popularity-contest/+bug/493852 "The goal of popularity-contest not only collect what packages are installed, but also determine what packages were recently used. In order to do so, It checks what files are currently open and scans last access time for files belonging to installed packages." | 09:49 |
| ubot5 | Ubuntu bug 493852 in popularity-contest (Ubuntu) "popularity-contest will stat() every file installed file or ~80k files on my system which grinds things to a halt" [Undecided,Confirmed] | 09:49 |
| ducasse | that's pretty horrible | 09:52 |
| akik | although that's a 5 year old report | 09:52 |
| akik | still open | 09:52 |
| akik | here's a newer one from 14.04 :) https://bugs.launchpad.net/ubuntu/+source/popularity-contest/+bug/1510996 | 09:56 |
| ubot5 | Ubuntu bug 1510996 in popularity-contest (Ubuntu) "Popularity-contest consumes 100% CPU" [Undecided,New] | 09:56 |
| TJ- | akik: That's how it works, popularity-contest is a Perl script that determines the most recently used (MRU) file in each package. It's triggered by a cron.daily job | 09:57 |
| TJ- | I'm pretty sure that'll need reworking for Ubuntu if the plan goes ahead | 09:57 |
| akik | i have little hope about this system helping development if bugs related to the popularity contest are kept open with seemingly no progress | 09:57 |
| akik | it'll just bring in heeps of data in which the developers will drown | 09:58 |
| TJ- | It's because it's a Debian package for Debian usage so it's not been touched much since we stopped using it actively about 8 years ago | 09:59 |
| TJ- | Looks like the stats graphs need regenerating... can't make out the dates and looks like the last time it was generated wa 2016 | 10:02 |
| akik | wouldn't it be enough to see which packages the user installs after the initial install? | 10:02 |
| akik | scanning the whole file system is idiotic | 10:03 |
| TJ- | That made sense when Debian created it in the early 2000s but not now | 10:03 |
| TJ- | when we maybe had 50 packages installed :) | 10:04 |
| TJ- | It'd be much better to simply attach a kprobe to the kernel's exec*() functions and log the path of the executed binaries | 10:04 |
| TJ- | And use the archive.ubuntu.com server logs to monitor the actual packages installed | 10:05 |
| akik | yea :) | 10:08 |
| akik | that package installation logs are _already_ there | 10:08 |
| TJ- | Have you looked at the popcon graph? it looks like someone's been spamming fake results at it: architecture names include "34d64", "Here", and "ZX_Spectrum" | 10:09 |
| akik | i looked into it yesterday | 10:09 |
| akik | saw that 2016 at the bottom | 10:09 |
| akik | i see | 10:09 |
| akik | some improvement needed | 10:10 |
| TJ- | I suspect no-one actually looked at what the package does and how it behaves. | 10:10 |
| TJ- | I'm going to see if I can put together a simple, no-impact, systemtap probe to record the executables | 10:12 |
| TJ- | We seem to already have the tooling in-place, try this please: "sudo execsnoop" | 10:15 |
| akik | not installed, no thanks | 10:16 |
| TJ- | Oh, it's just me because I've been doing kernel debugging. Darm | 10:16 |
| TJ- | it's much more lightweight that popcon though | 10:17 |
| akik | which package does that come from? | 10:19 |
| TJ- | perf-tools | 10:19 |
| TJ- | it uses the kernel FTRACE (function trace) mechanism to record all execve() calls | 10:20 |
| TJ- | I think popularity-contest could be improved dramatically. Currently it checks for file-access times for *every* file in every installed packaged using foreach ("$dpkg_db/$pkg.list", glob("$dpkg_db/$pkg:*.list")) | 10:21 |
| akik | :) | 10:22 |
| TJ- | I would think that could be changed to simply monitor the executable files from PATH and then identify which package contains each exectuable using 'dpkg -S /path/to/execeutable' | 10:22 |
| TJ- | And with a delay within the while() loop it wouldn't consume CPU | 10:23 |
| akik | dpkg -S is a bit intensive to be run on a per file basis | 10:25 |
| TJ- | Not if run on the server. On the client all it needs to do is report a summary count of how many times each executable was used | 10:33 |
| TJ- | The server could keep an in-memory hashmap which would make the lookup trivial | 10:34 |
| TJ- | If instead of submitting daily the submission only occurred when X bytes of data have been collected that would also randomise the load on the popcon server - currently the code tries to take server load into account | 10:35 |
| akik | just to make clear, i'm totally against this new data collection | 10:49 |
| BluesKaj | Hi folks | 11:09 |
| ducasse | morning, BluesKaj | 11:25 |
| BluesKaj | hey ducasse | 11:25 |
| BluesKaj | ducasse, my son and I are going to install anew mobo and an i5cpu on this pc> What's your experience with legacy mode on UEFI, will it boot a msdos table ? | 11:28 |
| BluesKaj | any ifo about such a setup would be helpful | 11:29 |
| BluesKaj | info | 11:29 |
| TJ- | BluesKaj: yes, it will | 11:30 |
| ducasse | i haven't really used legacy mode on uefi systems, i just install in uefi mode. | 11:30 |
| TJ- | BluesKaj: you can pre-test it with a custom USB key to prove it if you want to be 100% sure | 11:30 |
| BluesKaj | cool TJ- thanks ...I have all my valuable data backed up to 2 different computers, just to be safe | 11:34 |
| BluesKaj | valuable data being family pics, fav music and movies etc | 11:35 |
| ducasse | better safe than sorry | 11:36 |
| === DalekSec_ is now known as DalekSec | ||
| BluesKaj | yup | 11:39 |
| BluesKaj | anyway it should be an interesting exericise | 11:42 |
| ducasse | what kind of mobo did you get? | 11:44 |
| BluesKaj | my son got a good deal on a MSI Z370-A Pro | 11:47 |
| BluesKaj | plus an i5 8400 cpu, 8G crucial Ram,and aEVGA 600W PSU | 11:50 |
| BluesKaj | he wanted to get more Ram , but i thought 8G is plenty for my needs | 11:52 |
| ducasse | sounds good | 11:52 |
| BluesKaj | he's rebuilt several pcs ...this will be my first | 11:53 |
| ducasse | you'll have fun, then :) | 11:53 |
| BluesKaj | he says that MSI has received false unjust bad publicity about their products...he hasn't any problems with their mobos...ever | 11:55 |
| BluesKaj | yeah , it should be fun | 11:56 |
| ducasse | i've mostly used asus boards, no experience with msi myself | 11:56 |
| BluesKaj | right, we considered an asus board , but msi deal was too good to pass up | 11:57 |
| JimBuntu | I have had issues with MSI and Asus MBs. I would say that since I was a builder/maintainer for long enough that I have probably seen issues with all brands we ever purchased. Overall, I agree that MSI has taken a lot of what appears to be disproportionate hate from people buying the cheapest boards and expecting higher quality or longer life spans. | 12:24 |
| JimBuntu | While it's not always the case, a mobo that costs 4 times as much, at least had the finances behind it for the manufacturer to have used higher quality/fresher components, run more rigorous testing and could afford more QA on the built products. | 12:26 |
| JimBuntu | I had a batch of CPUs that we could only get to show their fault by using them in a flight simulator... vendor didn't want to exchange/return them so I had to drive the lot of them back to them and show them the issue and how they otherwise passed all of our testing. Once they saw it, no more hassle. | 12:27 |
| TJ- | I recall something similar with one of my guys discovering some weird issue and having to prove it to distributor; this was back around 2002 | 12:30 |
| JimBuntu | TJ-, this would have been around the same time, +/- a couple years. | 12:31 |
| TJ- | JimBuntu: be ironic if it were the same batch :D ... maybe you got ours, or we got yours! | 12:32 |
| JimBuntu | TJ-, lol. It's possible, or that we were getting them from the same manufacturing LOT. I'm sure their LOT sizes were bigger than our distributors. I'm trying to remember if it was a Pentium or an AMD K2 series, pretty sure it was a Pentium flavor. | 12:34 |
| JimBuntu | After a quick check, I'm fairly certain it was a K6-III as it wasn't a slot processor... oh do I remember how much $ I spent on my K7 Athlon | 12:38 |
| TJ- | Back then we were using AMD only, their CPUs were so much better. | 12:39 |
| TJ- | I've still got an Asus A7D266 with dual socket AMD Athlon 2000 MPs running as one of my legacy forensics servers | 12:39 |
| JimBuntu | I was always on the fence... it was always a tossup, spend more money for more reliability or go cheaper and spend that money on something else. I bought the Athlon on like day 2 that it was available though, so I paid quite a bit and it actually worked great, even wound up with the overclock device that could be slid onto it | 12:40 |
| JimBuntu | If I dig around enough, I should still have a dual Pentium Pro machine laying around somewhere. I don't leave these older beasts running, they simply hold the floor down for me. | 12:43 |
| TJ- | or just a graphite pencil rubbed over the traces :) | 12:43 |
| JimBuntu | TJ-, yeah? I had never heard of doing that for the Athlon or at least I don't remember. THe add-on board was pretty inexpensive | 12:44 |
| TJ- | here it is! http://computer-communication.blogspot.co.uk/2007/06/unlocking-duron-and-athlon-using-pencil_08.html | 12:46 |
| TJ- | basically joining exposed links | 12:46 |
| JimBuntu | That's hilarious! | 12:48 |
| TJ- | That's true hacking :) | 12:49 |
| JimBuntu | I know I still have the Athlon itself in it's custom Aluminium case... unsure if I kept the mb | 12:49 |
| === leftyfb_ is now known as leftyfb | ||
| === acheronuk_ is now known as acheronuk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!