Root Access

Meh, I don't use my router for WLAN.

;p

Also being a total PITA to upgrade was the main reason I actually switched back away from x86 routers and back to just plain Openwrt on a plain hardware router.

@qasdfdsaq neither, but my APs are an assortment of consumer routers configured to act as pure APs

3:04 PM

I really build my network based on typical enterprise principles. Switches are separate switches. Routers do routing and nothing else. Wireless APs are wireless APs.

I should probably try Ubiquiti sometime

Hence wireless coverage doesn't come from the router but a pile of ~~strategically placed~~ out-of-reach-of-cats wireless APs.

looks at prices

on second thought let's not

Ubiquity are relatively cheap compared to... well Cisco.

The controller-less WAPs and bridges are actually priced below many consumer routers.

wifi-stock.co.uk/details/ubiquiti-powerbeam-pbe-5ac-400.html

ubiqity's WAPs are what I'd go with if I was buidling out my own home wifi

used them at my last job, they're really nice.

3:16 PM

I wouldn't at home. I don't see any real advantage in a home router environment.

They have great long-range PtMP protocols on the outdoor bridges. On the home routers? It just any old Atheros router with slightly locked-down Openwrt.

@qasdfdsaq I actually need multiple APs in my current apartment

Then you're even better off with just plain Openwrt.

@qasdfdsaq the nice thing with the regular/short range ubiquities is that you can manage them all at once

That's... why I use openwrt.

Do the new Ubiquiti routers have cloud management now? (i.e. you don't need to buy an on-site controller)?

Coldplay went from youtube.com/watch?v=yKNxeF4KMsY => youtube.com/watch?v=YykjpeuMNEk :(

3:22 PM

> UniFied Wi-Fi and Public Address System Integration
Make announcements over Wi-Fi — no separate hardware needed.

Mobile Announcement App
Broadcast announcements with clarity from your iOS or Android-based device using our mobile app.

Coming soon to the App Store and Google Play.

LOL, there's something I didn't know I was missing...

0_0

I COULD TOTALLY USE THAT.

LIKE SERIOUSLY.

@qasdfdsaq I think they've had that a while

Why would a dog need a public address system in a wireless AP?

@qasdfdsaq my apartment is wierdly built

and my mom yells for me across the apartment.Which 1. I cannot always hear. 2. Annoys me. 3. Annoys here.

I'd just throw on the app and let her use that.

LOL nice.

o.O

Scanning

Creating archive originalsystem.img.7z

Compressing  originalsystem.img

Everything is Ok

3:27 PM

actually that's also why my wifi is that bad

466G originalsystem.img
253G originalsystem.img.7z

That's still disappointingly big :S

I wonder what the hell it actually is.

there's a section where its a narrow corridor between two halves of the apartment

Nature of data? Low compression settings?

@qasdfdsaq I assume it's a disk image.

kpartx?

3:29 PM

It's just a giant image I found on my old server -_-

I'm trying to figure out what I haven't migrated yet, and take a full backup.

Gotta delete some of this crap first though.

Perhaps it has backup images on it :D

I suspect I won't need any of this.

875M blog
2.1G clio
1.2G firefox
53G leto
119G phoebe
1.8G terraria
574M web

Let's see.

blog is long outdated and empty iirc.

112G phoebe

clio was a gitlab instance and has already been replaced

53G leto

uhuh

3:31 PM

firefox is ff-sync and has been replaced/migrated

leto is downloads and why the hell is it so big

phoebe is experimental and why the hell is it so big

terraria is unused and the worlds are stored on dropbox anyway, though maybe I should back up the configs

web is just a reverse proxy... should probably grab my ssl keys

@Bob That's what she said

51G leto/root

41G leto/root/CS_Servers
9.7G leto/root/CS_Servers.7z

ahhhhh

See I was right ;p

lol

@HackToHell that's not the img from earlier

36M phoebe/chatbot

o.O

3:33 PM

41G leto/root/CS_Servers
41G leto/root/CS_Servers.tar
36G leto/root/CS_Servers.tar.gz
9.7G leto/root/CS_Servers.tar.gz.7z
10.3G leto/root/CS_Servers.tar.gz.7z.rar

112G phoebe/home/bob

107G phoebe/home/bob/test.out

...what on earth

lol

112G phoebe/home/bob
112G phoebe/home/bob.tar.tmp
112G phoebe/home/bob.tar.out
3G phoebe/home/bob.tar.xz
112G phoebe/home/bob.tar.xz.rar

cat that file plz

for scienece

698M phoebe/home/bob/goldfish

phoebe/home/bob/test.out: ASCII text

nope.

2

oh god you're kidding

3:35 PM

hahaha

why the hell do I have that

Can you try and compress that

;p

MMMLMMMLA
MMMLMMMLB
MMMLMMMLC
MMMLMMMLD
MMMLMMMLE
MMMLMMMLF
MMMLMMMLG
MMMLMMMLH
MMMLMMMLI
MMMLMMMLJ
MMMLMMMLK
MMMLMMMLL
MMMLMMMLM
Calculating line numbers... (interrupt to abort)

And I get kicked for counting three numbers -_-

@Bob wc -l

3:36 PM

@HackToHell You realise it's going to take a while to read a 100 GB file from a HDD?

screw it

@Bob It would take one minute on my ZFS array.

let's try this

How long would it take to upload it to my ZFS array?

root@gaia:/var/vserver/phoebe/home/bob# 7za a -t7z test.7z test.out -m0=lzma2 -mx=24

@qasdfdsaq this is on a single raw hdd. old server.

new one is raidz1 on three SSDs and would be considerably faster

My raidZ is on over 9000

3:39 PM

yes, I am running it with 24 threads

hm... wonder if 12 would be faster

nah, there's I/O happening

> The risk of sexual transmission of Zika virus is thought to be very low. However, sexual transmission has been reported and so if a female partner is at risk of getting pregnant, or is already pregnant, condom use is advised for a male partner arriving from an affected area for the following durations:
 28 days after his return from a Zika virus transmission area if he has not had any symptoms compatible with Zika virus infection
 6 months following recovery if he has had symptoms compatible with Zika virus infection.

Thanks boss, I really needed work to tell me how to not get pregnant.

Because, er... Northern Scotland is at such high risk...

> OVH cannot be held liable for technical incidents on the backup spaces or the data hosted.

...uhm

isn't the whole point of backup, I dunno, reliability?

dammit, @HackToHell, I wanted to get things done, not spend an hour watching this compress -_-

3:57 PM

@Bob :D But it's .... fun

4:23 PM

"Netflix and chill"

@Bob

This made me think of you

big gif

Wait you hate huge GIFs right? Oops :-(

lol

you can still edit it to a gifv :P

I don't mind so much when I'm not on mobile, but it's probably better to edit anyway.

You have ... ~1min left to edit before

Yeah but I don't want people to have to clikc it

Because then nbody will clikc it

And nobody will see it

@qasdfdsaq ...it's 75 MB :S

Oh that's worse than I thought

I hope it's not on the first page of chat tomorrow

4:59 PM

We have some time to make sure it won't be.

Ssshhh! Nobody say anything

It's so cute. Lets keep it here

I can always remove it.

20 secs ago, by qasdfdsaq

It's so cute. Lets keep it here

Removed.

:-(

So much for democracy

5:00 PM

Who said this site was democratic?

Isn't it from the land of freedom and liberty that likes to thrust it's own brand of democracy on the rest of the world?

And aren't moderators meant to be democratically elected?

5:15 PM

hmm

kvm doesn't have automatic memory ballooning :(

5:28 PM

Does it have manual memory ballooning?

linux-kvm.org/page/Projects/auto-ballooning

Seems to... at least partly exist

@qasdfdsaq Yea, I was reading that.

But manual is useless.

Screw it, created a 32 GB swap on ZFS -_-

now I just need to figure out why I'm not getting an IP...

5:49 PM

o/

6:02 PM

ugh

I broke my server :S

:(

I have a planned server breakage upcoming in May... a week or so after Ubuntu Server 16.04 LTS is released, I'll upgrade to it and break LXD and have to re-learn how to use LXD 2.0 without lxc.raw

should be able to finally install Fedora or RHEL in a container with 16.04 running on systemd

The following packages have unmet dependencies:
 libguestfs-tools : Depends: libncurses5 (>= 6) but 5.9+20140913-1+b1 is to be installed
                    Depends: libtinfo5 (>= 6) but 5.9+20140913-1+b1 is to be installed
                    Depends: libguestfs-perl (= 1:1.30.6-1+b1) but it is not going to be installed
E: Unable to correct problems, you have held broken packages.

welp

I need the testing version because the stable version is bugged

but here I come to dependency hell

I had that same problem, trying to think what I did to fix it

I think I compiled libguestfs, which was a bastard because it depends on the world

apt-get -s --no-install-recommends install libguestfs-tools/testing libncurses5/testing libtinfo5/testing libguestfs-perl/testing

The following packages have unmet dependencies:
 libguestfs-perl : Depends: libguestfs0 (= 1:1.30.6-1+b1) but it is not going to be installed
                   Depends: libxml-xpath-perl but it is not going to be installed
                   Depends: libwin-hivex-perl but it is not going to be installed
                   Depends: libsys-virt-perl but it is not going to be installed
 libguestfs-tools : Depends: libguestfs0 (= 1:1.30.6-1+b1) but it is not going to be installed
                    Depends: libsys-virt-perl but it is not going to be installed

NOPE

which tool do you need from libguestfs?

6:11 PM

@allquixotic virt-builder

I was using it before.

Made the stupid mistake of trying to update

...boom.

come to think of it, I should have a rootfs snapshot from before the breakage

brb praying to the zfs gods

hmm, I have a working copy of libguestfs-1.30.4 on my box

The following packages have been kept back:
  libguestfs-perl libguestfs-tools libguestfs0 perl perl-base

\o/

zfs rollback tank/ROOT/debian-8@preupgrade

that was a lifesaver

libguestfs-tools:
  Installed: 1:1.28.12-1
  Candidate: 1:1.30.6-1+b1
  Version table:
     1:1.30.6-1+b1 0
        750 ftp.us.debian.org/debian testing/main amd64 Packages
 *** 1:1.28.12-1 0
        100 /var/lib/dpkg/status
     1:1.28.1-1 0
        995 ftp.us.debian.org/debian stable/main amd64 Packages

bugs.debian.org/cgi-bin/bugreport.cgi?bug=777548

They fixed the bug in 1.28.6-1

how costly in disk space are snapshots?

@allquixotic it's copy-on-write

I took a snapshot JUST before I went down that rabbit hole.

should i take a snapshot of my entire tank? :P

things are working good now

6:15 PM

@allquixotic I'm going to set up proper backups (zfs send | xz | (some encryption)) to OVh's backup store sometime...

nice :D

ovh doesn't give me enough data in the backup store for me to do that

475G alloc

55.6G alloc here

> On the OVH dedicated server range, you get a backup space of 500 GB per server. In order to offer you a quality service without the hassle of technical administration, this space is entirely managed by OVH.

wait why is it 832G total o.O

so if I deleted a bunch of old crap I could probably do it

6:18 PM

@allquixotic you can pay some $12/month for 1 TB

$16/month

which is ouch

ah

yea, could probably ship it off to s3 :P

or glacier

Lightly Salted

Greetings

Is this chatroom tied to a IRC channel?

@allquixotic Glacier would cost you $3.50/month for 500 GB

@LightlySalted nope

Glacier restores are considerably more costly, since you get charged for the data transfer out AND for glacier restore

but I suppose that's basically worst-case.

technically I could use Crashplan

Q: What does the (!) mean after uptime on htop

6:31 PM

11

I was analyzing some web heads looking at htop and noticed the following Uptime: 301 days(!), 23:47:39 What does the (!) mean?

htop

lol

heh. I think my server uptimes have been getting progressively longer over the years as laziness and general fatigue about messing with the internals of my systems sets in; I only reboot for major dist upgrades and such (or to move from one physical server to another)

58 days uptime now, but I've had a few 3-digit uptimes (in days) and this one will hit 100 days before I reboot for 16.04 in May

@allquixotic Mine is over 150 now :\

I should probably reboot and let the kernel updates apply

kernelcare ftw

root@debian:~# uptime
 18:51:48 up 210 days,  6:43,  1 user,  load average: 0.20, 0.22, 0.40

My WI server is only 60 days but that's cause they keep having power issues :\

I've never had a load average below 1.00 ever since I started using zfs, and I'm not sure what's doing it, but it doesn't seem to negatively affect server performance

it used to be a lot worse before I killed about 5 things that were eating a core apiece

6:39 PM

@allquixotic Does that work with the custom zfs builds?

Hm... I guess zfs doesn't touch the kernel, just the initramfs

@Bob it doesn't care if you have custom modules; it doesn't patch your custom modules, but it doesn't change the semantics of core kernel interfaces anyway, so zfs continues to interface with the core kernel normally

load average: 0.07, 0.17, 0.36

as long as the core kernel binary is unmodified it will work, and I think it also patches modules that are part of the default distro

means I have to reboot to get a new ZFS version though, which is what I did last time I rebooted

can't remove the module while it's my root fs :P

osv.io

o.o first time seeing that

6:56 PM

@Bob I think most of my usage comes from running Windows 10 in kvm-qemu :P

7:18 PM

3.7 GB of FF sources

7:30 PM

Baw, I don't have htop
top - 19:29:46 up 523 days, 1:21, 2 users, load average: 0.00, 0.00, 0.00

Wait 2 users?!

o_0

Also I have no idea what my password is :-/

Is it qazwsx ?

Nope

I don't think any password of mine has ever contained 'q'

Or 'a' for that matter

Or 'z'

12345 then . ;-)

Or 'w' or 's' or 'x'

Dammit what does my password contain?!

Worst case: su to toor and passwd yourusername

7:40 PM

The only username allowed on the system is 'root'

(Thanks Xenserver...)

My root user is called toor (and root is another uid 0 user)

Xenserver doesn't really support user accounts like normal linuxes

Xen as controlling domain?

Citrix Xenserver as dom0

<--- Not touched Xen yet, though their devs at FOSDEM make it look really nice

Ah

7:42 PM

Xenserver != Xen Server

The demo has FreeBSD as dom0

Xenserver is a citrix proprietary thing

OK. New to me.

Based off RHEL 5 or something

I planned to look into xen and esxi etc

Got a nice new desktop to play with

7:43 PM

Xen is the underlying hypervisor.

Xenserver is basically a proprietary distro, management tools, and GUI that Citrix shipped with it

But as yet I have no installed any type-1 hypervisors yet.

Xenserver is also based on a really old version of the Xen hypervisor

Why do I even call them Citrix. Shit-trix was so obvious.

I have a habit of substituting "Shit" into any company name where it fits.

Cisco = Shitsco is my favourite

Microshit server 2008 ?

Yup, MicroShit

Shitrix

I also like renaming delivery companies. Royal Fail. Parcel Farce. TNTits. DHfaiL. ShittyLink

UPShit - what can brown do for you?

7:52 PM

Oddly enough, I have never had major, systematic issues with any of the courier services here.

Today I collected my SSD, shipped by PostNLshit with 'Delivery to destiny home only' from my neighbours

We generally prefer to ship using USPS Priority Mail, which in our experience is fast yet affordable.

I've had issues, but they're rare enough that it's fine

One particular company explicitly lies

sigh Busted battery access door on my Metz mecablitz 58 AF-1 digital flashgun. I paid $400 for this flashgun and the battery door is so poorly designed.

I guess I'll need some gaffer tape to hold it down.

Granted, I don't use the flash often but when I do need it the battery door becomes a serious issue. Regular tape just isn't strong enough to hold it down for more than a few minutes.

There's a reason professional photographers usually carry a roll of gaffer tape with them...

7:59 PM

There is, but I always assumed it was for when they dealt with children.

Next time I get to B&H, I'll grab a couple of rolls of 1" gaffer tape and keep one in my camera bag at all times.

It's not exactly cheap, but not too expensive, either.

8:35 PM

Is it normal for an Android phone (with the cellular radio on) to experience extremely high battery drain when the SIM is removed?

Yes and no.

Depends what network it's on

AT&T. Battery noticeably drained after removing the SIM for about 30 minutes.

The phone indicated red for the cellular network signal in the battery graph for the time during which no SIM was inserted.

I thought the cellular radio turns off if the SIM is removed, switching on only if you try to make an emergency call.

I don't really know much about the U.S. network configs, and certainly nothing about the LTE ones.

What it indicates in the notification bar is manufacturer proprietary

In general a phone without a SIM will constantly try to search and register on every network it sees so it's already registered in case you try to make an emergency call.

> No SIM card - Emergency calls only

Basically it acts the same way as if you had a SIM card in, but your operator didn't have coverage in the area.

8:41 PM

Hmm...

Emergency calls only would imply it's parked on another network, or at least attempting to.

That might make sense.

(Otherwse it tends to say "No service")

Speaking of phone battery life, my iPhone 6S Plus is eating battery faster than an overclocked HTC Thunderbolt in my office @_@ workload is bluetooth listening to music with occasional low-bandwidth downloads over LTE

> No SIM card - No service

8:42 PM

at home with good signal, I used 10% battery life in a full working day while working from home

at office with crap signal, LTE extremely marginal, I'm down to 66% after 5 hours

I know in the UK they used to allow emergency calls without a SIM, then they didn't, then they did, then they didn't again.

I have a fair to good LTE signal here at home. Good enough for reliable calling and texting but data speeds can be spotty at times.

Battery only drains about 5-10% overnight.

@allquixotic There's a lot to it than just signal though

It's mostly up to the network config.

When and where a phone "searches for a better signal" is defined purely by the network.

I try not to run too many apps that consume battery power in the background. Android Doze does a good job keeping power usage under control.

I do have OpenSignal installed on my phone.

The conditions under which it switches between LTE, HSPA, GSM, etc. and whatever else you Americans have, is also configured by the network

8:45 PM

(Nexus 5X, 32GB, running CyanogenMod 13 nightlies.)

LTE signal tends to be quite widely under-reported on Android handsets too

Hmm, OpenSignal ... might have an effect

I dunno about on Android 6, but on Android 4.4-5 there was a "feature/bug" that meant measuring signal level required you to have your screen on

So apps like OpenSignal, Sensorly, etc. would have to turn your screen on to take a signal measurement. Wanting to log in the background meant keeping your screen on 100% of the time.

When I open the OpenSignal app, the cellular handoff behavior seems to change a bit.

I thought OpenSignal logs signal data in the background?

It's configured for "Medium" rate data collection.

On Android 4.4-5.0 it can only do that if the screen is on.

Doesn't seem to be a big battery hog.

Then again, I don't see any wakelocks associated with OpenSignal.

It possibly only logs when the screen is already on.

I know Sensorly keeps a full screen-on wakelock if you try to turn the screen off while logging.

8:49 PM

Okay. I see more than 100 wakelocks from OpenSignal using the Wakelock Detector app (requires root) over the last 28 hours but it doesn't keep them for long. The effect on standby battery life is negligible.

Probably periodic logging then. Turns the screen on every 15 minutes to get a reading, then turns it off again.

I think it's also theoretically possible for an app to be notified by the system when cell ID changes, rather than having to constantly poll for changes

@qasdfdsaq ...more accurately, takes the CPU off the idle state to collect and submit data.

As I said, it can't collect data with the screen off whatever state the CPU is in.

This seems to be a very quick process. The effect on the battery is probably less than 0.5% in any case.

@qasdfdsaq If it takes a partial wakelock, it should be able to do stuff in the background without turning on the screen.

It's nothing to do with wake lock.

Android simply doesn't provide updated signal readings with the screen off.

They probably thought it was a good idea to save battery... except it results in the opposite

8:54 PM

Not sure if I can say that. Variations in signal strength are reported in the battery graph even when the screen is off and the device is not awake.

Yeah, that's the OS. It doesn't have to use public APIs

code.google.com/p/android/issues/detail?id=10931

Well that's odd

If you're looking for a phone with extreme battery life (at the expense of a very slow processor and poor camera), you might want to consider the CUBOT H1.

anandtech.com/show/9868/cubot-h1-smartphone-test-5200-mah