- WLUG - wlug.mailman3.com

Next Meeing Oct 11th, 2018! NOTE LOCATION: RC61! Topic: TBD
by Tim Keller 08 Oct '18

08 Oct '18

Hey Gang, We've got a meeting! Date: Oct 11th, 2018 Time: 7pm Location: WPI Recreation Center Room 61 (RC61) Topic: TBD If anybody has a topic burning a hole in their pocket let me know. I've already got next months lined up! One option was to possibly demo a raspberry pi running a true 64bit version of debian. As usual, snacks and refreshments will be provided and afterwards we'll head off for dinner to continue the conversation. Thanks, Tim. -- I am leery of the allegiances of any politician who refers to their constituents as "consumers".

1 0

Charter Telephone reviews, boylston?
by John Stoffel 29 Sep '18

29 Sep '18

Hi, Anyone on here using Charter Telephone VOIP service? I've currently got Verizon, but the wife hates it because our phone lines get flaky all the time, esp when it rains. So I'm thinking to save money and combine all my stuff onto Charter. I've already got High Speed internet and regular old cable. Not wild about Digital Cable since I'm happy with Tivo and I don't want yet another set top box to have to deal with... So, any horror stories about Charter Phone VOIP quality and service? Thanks, John

16 34

Reminder! Meeting Tomorrow! Sept. 13th, 2018 @7pm.
by Tim Keller 12 Sep '18

12 Sep '18

Hey Everybody, This is a reminder we've got a meeting on Thursday Sept. 13th, 2018. Time: 7pm Location: WPI Campus Center Mid Century Rm (rm 331) "The usual spot" Topic: git LFS (Large File Support) I thought I would give a quick talk using git as a backup solution for larger files. After that I figured the conversation would branch out. I'd also like to poll the group and see what people would be interested in for topics. As usual snacks and refreshments will be provided and we'll head off to dinner afterwards to continue the conversation. Thanks, Tim. -- I am leery of the allegiances of any politician who refers to their constituents as "consumers".

1 0

Next Meeting Sept. 13th at 7pm Location: WPI Campus Center Mid Century Room (rm 331)
by Tim Keller 07 Sep '18

07 Sep '18

Hey Everybody! Our next meeting is going to be on September 13th at 7pm. Location: Campus Center Rm 331 (Mid Century Room) Topic: Right now I don't have a main topic, but if people would like I can do a quick talk on my current backup system that's using git / gitlab and offsite repo pushes. As usual, snacks and refreshments will be provided and we'll head off for dinner afterwards. Thanks, Tim. -- I am leery of the allegiances of any politician who refers to their constituents as "consumers".

1 0

Software packaging/deployment
by Richard Klein 30 Aug '18

30 Aug '18

I've brought up Snaps before (https://snapcraft.io/) If those interest you, you may want to check out Flatpak (https://www.flatpak.org/) also. I haven't tried it out, yet, but it's on my to do list. -- Rich

2 1

Re: Monitor Madness
by Aljerin Butler,Jr. 17 Aug '18

17 Aug '18

Try doing some research here. | | | | Howtos, tips, tutorials; IT ebooks Super User | | | On Wednesday, August 8, 2018, 9:53:12 PM EDT, Krichevsky, Nicholas Jacob <njkrichevsky(a)wpi.edu> wrote: Hey all - By an interesting turn of events, my fellow interns and I were given the opportunity to build a computer to theoretically drive 192 monitors. Yes, really. We're running into a bit of a brick wall, so I figured I'd reach out and see if anyone had any crazy ideas that just might work. This project started because of a certain obsession of our CEO to drive as many monitors off one PC as possible. We put together a build using8x Radeon 7870s with 6x Mini DisplayPort outputs (the price spiked after we bought these for some reason) with2x Intel Xeon E5-2630 V4 , all tied together in a SuperMicro 4028GR-TR2. Limiting the cards to PCIE 8x, we should be able to drive all 8 with 16 PCIE lanes to spare. Because the Mini DisplayPorts follow the DisplayPort 2.1 spec, we were hoping to use MST with48 of these MST hubs from Monoprice (which we actually got on sale!). This would allow us to drive four displays @ 1080p off one Mini DisplayPort, allowing us to drive 24 monitors off one card, across 8 cards, totaling 192 monitors. With this all in mind, and the knowledge that we may have to write some funky Xorg configs, we pulled the trigger on all of it, and would attempt to drive as many monitors as we had on hand in the office. We decided to load Fedora on this machine, which should be able to drive MST with the included amdgpu driver, thanks to patches that were added to kernel 4.15, and some DC display code settings that were enabled in kernel 4.17. With all of this, we were able to successfully drive ... two displays off each MST hub, instead of the expected four. Not as fruitful as we may have hoped, but hey, 96 monitors would still be awesome. Out of sheer curiosity, we threw Windows on another SSD, and discovered that Windows could drive four displays off each MST hub, but we were getting hard limited to six monitors per card, no matter what we tried. We hypothesized that this might be due to theproprietary AMD driver setting an arbitrary limit, and thought we might be able to still pull off the 12 monitors per card in Linux. A couple of small scripts later, we were up and running with six monitors on one card. Upon adding the seventh and re-running our script, we were greeted with the following. [monitormadness@monitor-madness ~]$ python3 set-monitors.py Executing `xrandr -d :0 --output DisplayPort-3-1-1 --mode 1920x1080 --output DisplayPort-0-1-1 --mode 1920x1080 --output DisplayPort-1-1-1 --mode 1920x1080 --output DisplayPort-0-1-2 --mode 1920x1080 --output DisplayPort-3-1-1 --mode 1920x1080 --output DisplayPort-1-1-2 --mode 1920x1080 --output DisplayPort-4-1-1 --mode 1920x1080` xrandr: cannot find crtc for output DisplayPort-4-1-1 We ran xrandr --listproviders to see what we got. With two GPUs connected, we got the following. [monitormadness@monitor-madness ~]$ xrandr --listproviders -d :0Providers: number : 2Provider 0: id: 0x8d cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 6 outputs: 30 associated providers: 1 name:PITCAIRN @ pci:0000:04:00.0Provider 1: id: 0x4b cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 6 outputs: 6 associated providers: 1 name:PITCAIRN @ pci:0000:06:00.0 Looks like we can only get 6 CRTCs per card. This is where I reach out to you all. Does anyone know of some Xorg or kernel flag trickery we can use to get more outputs for our buck with the limited number of CRTCs? I don't quite know enough about how video cards work to make an informed decision about what's going on, but my hypothesis is that the card's firmware is what transmits how many outputs we can use, so we may be SOL in the department of getting more CRTCs, but perhaps there may be some way to trick Xorg into rendering the content of one CRTC across multiple monitors? Not sure. Would love to hear what you all think. Thanks for taking the time to read this! :) _______________________________________________ WLUG mailing list -- wlug(a)lists.wlug.org To unsubscribe send an email to wlug-leave(a)lists.wlug.org

1 0

Reconnecting with my media drive
by Richard Klein 09 Aug '18

09 Aug '18

When I was running Debian, I had most of my media segregated on a Seagate SSHD. That's still in the system now that I'm running Mint. I'm trying to get Plex up and running again. The file manager shows the media drive mounted, but shows nothing on it. I don't see it listed in fstab. I don't remember how it was formatted, but I think it's an ext4 drive. What should my next step be? -- Rich

2 2

Monitor Madness
by Krichevsky, Nicholas Jacob 09 Aug '18

09 Aug '18

Hey all - By an interesting turn of events, my fellow interns and I were given the opportunity to build a computer to theoretically drive 192 monitors. Yes, really. We're running into a bit of a brick wall, so I figured I'd reach out and see if anyone had any crazy ideas that just might work. This project started because of a certain obsession of our CEO to drive as many monitors off one PC as possible. We put together a build using 8x Radeon 7870s with 6x Mini DisplayPort outputs<https://www.newegg.com/Product/Product.aspx?Item=9SIA0AJ3BU1115&cm_re=Visio…> (the price spiked after we bought these for some reason) with 2x Intel Xeon E5-2630 V4 <https://www.newegg.com/Product/Product.aspx?Item=N82E16819117630> , all tied together in a SuperMicro 4028GR-TR2<https://www.supermicro.com/products/system/4u/4028/SYS-4028GR-TR2.cfm>. Limiting the cards to PCIE 8x, we should be able to drive all 8 with 16 PCIE lanes to spare. Because the Mini DisplayPorts follow the DisplayPort 2.1 spec, we were hoping to use MST with 48 of these MST hubs from Monoprice<https://www.monoprice.com/product?c_id=104&cp_id=10428&cs_id=1042802&p_id=2…> (which we actually got on sale!). This would allow us to drive four displays @ 1080p off one Mini DisplayPort, allowing us to drive 24 monitors off one card, across 8 cards, totaling 192 monitors. With this all in mind, and the knowledge that we may have to write some funky Xorg configs, we pulled the trigger on all of it, and would attempt to drive as many monitors as we had on hand in the office. We decided to load Fedora on this machine, which should be able to drive MST with the included amdgpu driver, thanks to patches that were added to kernel 4.15, and some DC display code settings that were enabled in kernel 4.17. With all of this, we were able to successfully drive ... two displays off each MST hub, instead of the expected four. Not as fruitful as we may have hoped, but hey, 96 monitors would still be awesome. Out of sheer curiosity, we threw Windows on another SSD, and discovered that Windows could drive four displays off each MST hub, but we were getting hard limited to six monitors per card, no matter what we tried. We hypothesized that this might be due to the proprietary AMD driver setting an arbitrary limit, and thought we might be able to still pull off the 12 monitors per card in Linux. A couple of small scripts later, we were up and running with six monitors on one card. Upon adding the seventh and re-running our script, we were greeted with the following. [monitormadness@monitor-madness ~]$ python3 set-monitors.py Executing `xrandr -d :0 --output DisplayPort-3-1-1 --mode 1920x1080 --output DisplayPort-0-1-1 --mode 1920x1080 --output DisplayPort-1-1-1 --mode 1920x1080 --output DisplayPort-0-1-2 --mode 1920x1080 --output DisplayPort-3-1-1 --mode 1920x1080 --output DisplayPort-1-1-2 --mode 1920x1080 --output DisplayPort-4-1-1 --mode 1920x1080` xrandr: cannot find crtc for output DisplayPort-4-1-1 We ran xrandr --listproviders to see what we got. With two GPUs connected, we got the following. [monitormadness@monitor-madness ~]$ xrandr --listproviders -d :0 Providers: number : 2 Provider 0: id: 0x8d cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 6 outputs: 30 associated providers: 1 name:PITCAIRN @ pci:0000:04:00.0 Provider 1: id: 0x4b cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 6 outputs: 6 associated providers: 1 name:PITCAIRN @ pci:0000:06:00.0 Looks like we can only get 6 CRTCs per card. This is where I reach out to you all. Does anyone know of some Xorg or kernel flag trickery we can use to get more outputs for our buck with the limited number of CRTCs? I don't quite know enough about how video cards work to make an informed decision about what's going on, but my hypothesis is that the card's firmware is what transmits how many outputs we can use, so we may be SOL in the department of getting more CRTCs, but perhaps there may be some way to trick Xorg into rendering the content of one CRTC across multiple monitors? Not sure. Would love to hear what you all think. Thanks for taking the time to read this! :)

1 0

Reminder: WLUG Meeting August 9th, 2018! Topic: Bro
by Tim Keller 06 Aug '18

06 Aug '18

Hey Gang, We've got a meeting on Thursday! Time @ Location: 7pm @ WPI Mid Century Room (Rm 331) aka, the usual spot Topic: Bro Network Security Monitor Presenter: Eric Martin Veteran WLUG member Eric Martin will be giving a talk about "Bro" What's Bro? Bro is a FOSS framework and event engine for analyzing network traffic. First developed at LBNL in 1995 it is used on public, private and personal networks provide robust indepth network usage and threat analysis. Check out http://www.bro.org for more details. Eric's talk will be focused around using Bro to understand the flow of traffic around devices on our home network. As we add an ever increasing set of devices that have the words "IoT" and/or Smart in their names, it matters that we know what these devices are doing and who they're talking to. Thanks, Tim. -- I am leery of the allegiances of any politician who refers to their constituents as "consumers".

1 0

Desktop distro
by Richard Klein 06 Aug '18

06 Aug '18

I'm having trouble deciding which distro to install on my new SSD. It's the only desktop computer in the house, so it's used for web, email, etc, plus it's a Plex server. DistroWatch.com shows Manjaro being really popular right now, but I've never heard of it. It's based on Arch, which I've heard of, but never tried. I see that Arch uses Pacman for a package manager. I'm used to APT; I'm not sure if I'd like Pacman. So anyway, I thought I'd ask for suggestions. -- Rich

5 9