« first day (1364 days earlier)      last day (3903 days later) » 

17:00
No idea
I was more focused on getting my gear back online
@ChrisS @pauska It's not. It's completely custom. Their model is that they map remote memory/disks into memory space on each server. You don't access remote memory rDMA, you just access memory and in hardware that gets translated into a remote call over their bus.
@Jacob that implies they did get the data back?
@Jacob That's bad right? "Show me my data!" "We have no data."
17:03
is it at least plausible that this might have been "an issue with the API"?
@MichaelHampton I never assume anything I say over a wire is private anyways
(it does seem like a thin coverup for catastrophic data loss of some kind)
@FalconMomot Not going to lie, I don't have the exact email, but my boss's snip to me was "Basically, after the research and tests conducted it became apparent that the data will not be recoverable and an SLA will be applied to all affected users"
@Jacob ouch. it's pathetic that they are calling it an API failure.
@FalconMomot Yeah tell me about it
I don't even know how big they are, and I think I know why...
Oh, all their support is done by incompetent people who can't speak English that told me that it was normal for my drives to go readonly fs
17:11
@Jacob Does that mean refund? b/c that's cool, but doesn't cover a bit of time
This HP server is doing my bidding!!
=> ctrl all show config

Smart Array P410i in Slot 0 (Embedded)    (sn: 50014380135F1E70)

   array A (SAS, Unused Space: 1144280  MB)


      logicaldrive 1 (16.0 GB, RAID 1+0, Transforming, 6% complete)
      logicaldrive 2 (542.7 GB, RAID 1+0, Waiting for Transformation)

      physicaldrive 1I:1:1 (port 1I:box 1:bay 1, SAS, 300 GB, OK)
      physicaldrive 1I:1:2 (port 1I:box 1:bay 2, SAS, 300 GB, OK)
      physicaldrive 1I:1:3 (port 1I:box 1:bay 3, SAS, 300 GB, OK)
      physicaldrive 1I:1:4 (port 1I:box 1:bay 4, SAS, 300 GB, OK)
TRANSFORM!!
(okay, i'm weird)
2
Heh, we had an issue with time warner resulting in a partial phone outage ...we got a $400 credit because of their SLA
@Jacob That's a "going out of business" level catastrophe.
@MichaelHampton They lost ours for sure. I got a nice HP box to replace it.
@MikeyB Hrm... I'm not crazy about all that. Sounds like a cache coherency nightmare. Also, would either need to be very selective about what gets shared over the bus or the OS would have to be heavily modified to deal with shared resources.
@MikeyB What would you be interested in this tech for? I'm struggling to find an application.
17:17
1
Embedded Systems Programming and Design

Proposed Q&A site for professionals and enthusiasts of embedded hardware and software. Created in lieu of the pending closure of Embedded Programming and Design, this is a reboot.

Currently in definition.

@Jacob So, they don't even have backups? >_<
@freiheit As far as I'm aware
@ChrisS Yeah the way you configure it is by dedicating a portion of system ram to be "shared". So essentially as long as your application is already expecting to be working with shared memory you can probably modify it to work.
@ChrisS For HPC work that deals with a lot of cross-connect data (fluid dynamics) I could see it being useful.
It's interesting, but seems like a really cool solution that needs a problem to solve.
Hrm... Maybe, but if you're designing an app for HPC communications anyway seems like going with an existing tech would be easier, also avoid vendor lock-in, standard libraries (should be easier to find developers), and the list goes on...
But the interconnect and software is cheap: $500-$1000
17:21
@MikeyB Exactly what I was thinking.
That is really cheap... Now you've got me interested - if only a little.
Bob
Bob
@BigHomie Wait what o.O
@ewwhite Oh sheeet, ouch. Bad dog!
@BigHomie I hate Area51 - so many proposals (and supporters) of sites that are basically "I know nothing of this subject and want someone to help me"
I fail at sysadmin - I had to look up: Windows-Server-2012-How-to-Reboot-Your-Server
@ewwhite I had to do that too :(. You're not alone.
17:29
@ewwhite nah, a failing sysadmin would ask How do I reboot my Windows Server 2012 server? on main.
@BigHomie quickly deletes
@ewwhite start powershell, run Restart-Computer
Poop nano isn't installed, but I need to edit the networking config. Whatever will I do?!? Use vi? Like I know how to do that.
Shame on you @Yacub, you should know start+R shutdown -t 00 -r -f.
@Jacob Every time I find a computer with nano installed, I remove it immediately.
17:31
@BigHomie I had a professor from Bangladesh call me that.
@Jacob I remember
@MikeyB Leave me and my inferior text editor ALONE
Using nano is a great way to turn your configuration files into worthless trash.
2
@Jacob I've added it as an alias to your account
Which is why I do infact use vim.
17:32
hmm, an amazon fresh driver has gone awol with my food.
node 'default' {
    package { 'nano': ensure => purged, }
}
node 'default' {
    package { 'logwatch': ensure => purged, }
}
17:44
node 'default' {
    package { 'bash': ensure => purged, }
}
ksh4lyfe
On that note, I think I can use puppet to save myself a lot of trouble...
node 'default' {
    package { 'openssl': ensure => latest, }
}
exactly
I don't even *know* this cryptic language, but what I know is

` node 'default' {
package { 'java' : ensure => purged, }
} `
all friggin day.
Markdown fail
17:47
^
indeed.
Nope not even going to try to fix that...
the sad part is it's not markdown's fault
For extra fun (make sure all openssl-using services get restarted to use the new library version):
node 'default' {
package { 'openssl':
ensure => latest,
notify => Exec[reboot],
}
exec { 'reboot':
notifyonly => true,
}
}
@BigHomie Wrong! It's ALWAYS Markdown's fault!
nah man, I love markdown
I blame the chat devops.
and myself, for not brining it to their attention and just complaining about it like an end user.
17:51
Now what I need is some way for security updates to generate tickets in redmine...
@freiheit I have a package that's distributed as a zip file of scripts. I'm trying to think of a better way to distribute it via puppet
RPM?
@ewwhite Build it into an RPM (or .deb)
not sure I want to go through the effort of RPM
it's 6 files.
Use fpm, it's dead simple
17:54
@ewwhite unzip it into your puppet repo
@MikeyB which I did already
and it works.
but just thinking about future.
or maybe that's a dead-end
@ewwhite file { ''path/to/whatever.zip': source => 'puppet:///whatever.zip'} exec { 'cd whatever ; unzip /path/to/whatever.zip': require => [File[whatever.zip],...], creates => 'whatever is created';}
@ewwhite But, seriously, build into an RPM, so you can version and "latest" and all that shit. :)
Already done.
4 hours ago, by ewwhite
  # Files
  $auto_snapshot_src = '/root/zfs-auto-snapshot-master.zip'
  $auto_snapshot_cmd = '/sbin/zfs-auto-snapshot'
  $auto_snapshot_cron = '/etc/cron.d/zfs-auto-snapshot'

  file { $auto_snapshot_src:
    ensure  => present,
    source  => "puppet:///modules/zfs-auto-snapshot/zfs-auto-snapshot-master.zip",
    mode    => '0755',
    owner   => 'root',
    group   => 'root',
    notify  => Exec["unzip auto_snapshot_src"],
  }

  exec { "unzip auto_snapshot_src":
    command     => "/usr/bin/unzip -o $auto_snapshot_src",
@ewwhite I thought you unzipped it in the puppet scripts
I meant like:
@MikeyB I found an error in the Makefile, so I'm just going to distribute the files manually.
17:56
file { '/etc/zabbix/zabbix_agentd.conf.d/zabbix_agent_userparam_postgres.conf':
    ensure  => file,
    mode    => '644', owner => 'root', group => 'root',
    source  => "puppet:///modules/zabbix/zabbix_agent_userparam_postgres.conf",
    require => Package['zabbix-agent'],
    notify  => Service['zabbix-agent'],
}
I'm customizing it enough that I'm going to ditch the Make structure from the zip archive
Man it'd be sweet if you could specify a source like:
source => "puppet:///modules/zfs-auto-snapshot/zfs-auto-snapshot-master.zip/usr/local/zfs-auto-snapshot.py"
But I do like the notify->exec(unzip)… kinda. It's an OK kludge.
and fighting some other issues.
I just push to a git repo, and jenkins builds the RPM for me. Just ensure => latest and that's all.
17:58
@MikeyB It's the sort of terrible thing behind the scenes of this automatic oracle install I'm tweaking...
@freiheit oracle DB? or another product?
@dawud DB
@freiheit You probably know this repos, but anyways: github.com/biemond?tab=repositories
@dawud forge.puppetlabs.com/biemond/oradb + forge.puppetlabs.com/hajee/oracle + a bit of tweaking on my side to get everything right.
ah, cool
18:04
Ugh, now even ISIS is sending me spam. Blocks and blacklists domain on the mail server
We still get tons of junk mail even after blocking all non-US traffic...
posted on June 06, 2014 by ryan

I recently wrote a bunch of Powershell Cmdlets that were related to each other, so I decided to package them all up together as a Powershell module.  These particular Cmdlets were tightly integrated with Active Directory, so it made sense that they would often be run on a domain controller, or against a domain controller, usually by a domain administrator. First, to cre

18:24
https://github.com/remixz/dogescript
Good afternoon ladies.
18:39
3
Q: Delete SCCM Certificate from Command Line

Wes SayeedSo we have a situation where a contractor deployed about 200 Windows 7 computers that were cloned improperly. The SCCM cert was not cleaned off the reference machine before it was sysprepped. Now because of the duplicate certs, the SCCM console is getting crapped up with invalid device records al...

I hate Comcast, so I'm shopping around for some other Internet connection. One company is offering to replace my 16/3Mb connection with a T1 for just a few dollars more per month. I have to wonder what planet these people are living on that they think anyone is going to pay more for a T1 over 16/3 CM.
@ChrisS me you and @mossy should talk.
@ChrisS Well your reliability and SLA would certainly benefit with that change. :)
«
We are having an issue after moved the server at Q9.
The power light on the front of the server is not coming on at all and Ilo is not working.
Power connection and network connections are all good.
Request to open an IBM ticket for fix the problem as the server is down.
We are not able to bring up and need IBM technician help.
»

How about you freakin' hook up the power properly and then we'll talk, eh?
19:10
so bumpy :|
@ChrisS I was paying $72/month for 12 down until about 1 hour ago.
I have a 3 down plan now.
wompwomp
@ChrisS Didn't you get the free upgrade? Everyone who was on 16/3 should now be 25/5
At least on residential. Business is another story.
Shot in the dark: are any of you using the pyvmomi API to vCenter?
19:25
Anyone using Azure?
-2
Q: Our Server is down, need help to figure out what is going on

user224188Our server is down and we do not have a proper sys admin to figure out what the problem is. The company that hosts the server has tried to reboot it and they are telling me it is booting and then crashing. They have tried to boot it in rescue mode, but nothing seems to be working. They have instr...

For every 1 firm that does things right, I reckon there is at least 3 more of these.
@DanBig Only slightly. And not even for Windows.
@MatthewIfe I estimate your estimate is low. None of the small businesses I know even do backups.
I set one of them up with overnight backups, but they wouldn't leave machines in sleep mode for WOL to work. eventually, they lost everything
@MikeyB get in there!
@MatthewIfe already posted :D
19:37
Because a) I was bored and b) It give me a opportunity to lecture the dude I was contemplating just logging in and fixing the issue (if it was fixable) but since he had no information I could use to contact him, he can go pay you for the pleasure.
LOL I can't believe I got badges for this
Idle curiosity: How many of you publish SSHFP records for your systems?
@voretaq7 no sir, not for any of our domains.
@voretaq7 I do
Only because FreeIPA does it automatically
I just put them on for my personal box because I rotated keys (the old ones were a little short for my liking), but I don't know many people who publish them
and it doesn't make much sense without dnssec I guess
19:52
@MikeyB man that guy really has it in for @MichaelHampton
Don't you know, I'm evil
Actually, this is evil.
@MichaelHampton evil? did someone call me?
@Chopper3 Oh, no, someone just mentioned yesterday's screaming child.
@MichaelHampton yeah, thats despicable
> It could have just been a guy on the phone with his girlfriend trying to find a gift for his kid, they should check the tape to see who was actually in the spot and look for anyone there
seriously?
20:11
Nobody takes out a loaded gun in the toy aisle at Target and leaves it on the shelf. Nobody sane, anyway.
> Ohai babe, nothin' just walkin through target with my tool out. You know haters gon' hate.
> Say, what does jr want for his birthday?
> Hang on, Let me see how much it costs, it's way on the top shelf
@BigHomie Don't be ridiculous. He'd shoot it down or shoot out the shelf supports.
@freiheit yeah, I had a feeling somebody was going there, nevermind this
@freiheit heh, did you see the verb just below it
So NetworkMangler decided to inexplicably drop my default route. Thanks a LOT.
20:23
There no logs?
@MatthewIfe Haha, LOGS?
@MatthewIfe The logs, they show nothing.
Meh, doesnt produce route changes (least those it seemingly doesnt manager).
Actually the logs are voluminous and unhelpfully unenlightening.
20:29
@ewwhite dude, when do you sleep??
He could tell you, but then he'd have to kill you get one of his minions to kill you
@Jacob The lack of bike lanes, you mean?
@freiheit Yeah I can't go anywhere
@tombull89 you sure? cuz I thought they were the ones sleeping for him :p
20:39
I don't have a road bike so I can't travel that fast
to prevent getting dead
@Jacob First thing to do, support local organizations trying to fix that: vabike.org/about-vbf bikenorfolk.org tbarides.org
@Jacob Second thing, find and take a LAB-certified Traffic Skills 101 class.
@freiheit Y?
@Jacob Because smart lane positioning does more to keep you from getting dead than paint does.
@freiheit But Virginia already tells me where to be in the lane, I just don't like the idea of riding in 45mph zones that much
20:46
Vehicular cycling (also known as bicycle driving) is the practice of riding bicycles on roads in a manner that is in accordance with the principles for driving in traffic. The phrase vehicular cycling was coined by John Forester in the 1970s to characterize the style of cycling utilized in his native country, the United Kingdom, in contrast to the deferential-to-cars style of cycling and practices that he found to be typical in the United States. In his book Effective Cycling, Forester contends that "Cyclists fare best when they act and are treated as drivers of vehicles". Technique A ...
@freiheit Forester's system of cycling has drawn criticism as being unsuitable for anyone but skilled, strong riders, with his assessment of skill being based upon speed. Bruce Epperson claimed[22] that, in a 1978 article, Forester required cyclists to sustain a speed of 18 miles per hour (29 km/h), a feat achievable by only 3% of Americans. The claim was subsequently refuted by Forester.[23]
@Jacob You mean like this?
Yeah pretty much
@Jacob Quick quiz: does virginia law require you to ride on the shoulder of the road?
@freiheit as close as safe, and not while avoiding hazards
20:52
@Jacob You sort of avoided the question, but if you read the definitions the roadway that you're required to ride to the right side of doesn't even include the shoulder. You're allowed on the shoulder, but never required.
@freiheit What's your point?
@Jacob My real point is that you just might learn stuff in those classes that would help you be safer and more confident riding. Bike lanes will take years to happen, so you might as well learn to be safe and happy with the roads that actually exist now.
@Jacob Wow, apparently I've been taking for granted the amount of bike lanes in my area.
Often I prefer to ride on the road, because bike paths are too slow.
And full of dangerous obstacles.
@Jacob yeah @freiheit is right, I thought it common sense but had no idea it was the law not to ride w/ two earbuds in
bike or car, I see people do it all the time
no worries, when I ride up, they know.
21:05
@EEAA I may also be spoiled (but all the fun rides are off the lanes and paths)
@BigHomie :(
@ewwhite You get 4 hours of sleep every other Sunday?
feels like it
@freiheit I just took my new commuter bike and rolled over to the fried chicken and condom restaurant a few blocks away.
21:16
@freiheit I've read all the laws, but I agree I do need to take a maintenance and safety course.
Let's try to solve my actual issue though?
Well actually looks like I might actually have
@Jacob The issue being that providence is no fun to ride on?
@freiheit It's too small, people can't safely pass as there is a lot of opposing traffic. So no, I'd rather avoid it past kempsville.
@Jacob So add some distance in the twisty suburbs?
21:26
@freiheit No need to go that extreme actually. bit.ly/1hEk5YZ
I sidewalk it between old providence and lord dounmore
@Jacob sidewalks are scary! :)
@freiheit how so?
adopt the Idaho stop!
@Jacob Sidewalks reduce how visible you are and force you into interactions with driveways and intersections.
@RyJones Outside of Idaho there's a strict "don't ask, don't tell" policy around Idaho stops... ;)
@freiheit yeah, I know
21:41
@RyJones Of course, around here, most motorists do the Idaho stop too...
@freiheit I thought that was the California stop
he says before he clicks the link
@RyJones We call it the california roll
The is a maki-zushi, a kind of sushi roll, usually made inside-out, containing cucumber, crab meat or imitation crab, and avocado. In some countries it is made with mango or banana instead of avocado. Sometimes crab salad is substituted for the crab stick, and often the outer layer of rice (in an inside-out roll) is sprinkled with toasted sesame seeds, tobiko or masago. As one of the most popular styles of sushi in the US market, the California roll has been influential in sushi's global popularity and in inspiring sushi chefs around the world in creating their non-traditional fusion c...
@MichaelHampton Mango or banana as a substitute for avocado?!? Barbarians!
@MichaelHampton you've messed around a bit with rhel7 right?
21:45
I've never liked the california roll. I prefer actual fish in my sushi...
@MatthewIfe Yes, why?
@MichaelHampton that is a proper California roll, yes
I'm unable to get systemd user sessions working, know anything about it?
@MatthewIfe Um, no? In what context?
Lets say some user wants to run an apache instance on some non-reserved port.
You're doing containers?
21:48
In F20, when you login you get a systemd user instance spawned automatically. It doesn't come up in RHEL7.
Last I heard, the LXC stuff was incomplete in RHEL 7
No, no containers.
@MichaelHampton I vastly prefer the rock and roll over the california roll. Or, even better, an alaska roll...
@MatthewIfe Well, I really don't want to run an apache instance at all...
Neither do I, but perhaps a user does.
Perhaps a user has written their own service they want to be persistent but benefit from being restarted if it stops, or other behaviours systemd supports (like bringing up dependencies)
21:52
Did I hear "LXC"?
in F20, if you do systemd status user-<uid>.slice you can see the systermd instance started. In RHEL7 it doesn't happen. I've looked through the source regarding how it is meant to be started in F20 -- but its all pretty much handled indirectly via dbus which doesn't help me get an 'easy' answer.
Only "HP", "Puppet" and "LXC" are enough interrupt my Tinder sessions long enough for me to log back into SF Chat.
In RHE7 I made some progress by copying the [email protected] file from F20 (which isn't in rhel7) but it still wont startup an instance on login automatically.
@ewwhite I've got this one puppet run up to 2974.41 seconds. I'm aiming for 86400 seconds.
@freiheit Man, mine run in 14 seconds or less! What are you making it do?
21:55
@ewwhite EVERYTHING
@ewwhite Also, that's just the initial run on system install. After that it's 30 seconds.
Oh, the things I do for you people...
@ewwhite (it's creating, resizing and mounting various LVM things, downloading a few gigabytes of oracle, unzipping and installing and patching the oracle stuff, then creating a database, starting all those things, creating oracle tablespaces and users...)
@MichaelHampton :)
@MichaelHampton Is that what the RHEL7 installer looks like for you? Mine's all kickstarted. :)
@freiheit hot!
21:57
@freiheit Yes, I'm doing a manual installation
And besides, if you kickstart you miss all the banner ads Red Hat throws in. Like this one:
@MichaelHampton I probably do one or two manual installs... then go back to kickstart
Yeah, I saw an ad for AdultFriendFinder.com during one of my graphical RHEL installations. That was weird.
7
Well, for some reason the DVD came up as RHEL 7.0 ComputeNode..so I had to improvise.
(and they're an HP shop.)
21:59
@MichaelHampton So, when do you think RHEL7 will actually get released?
@freiheit August
@MichaelHampton Not Tuesday?
@freiheit No, of course not. But they might announce the release date then.
@MatthewIfe Why wouldn't the user put that into crontab as a @reboot line and execute it via runit or something?
@MikeyB They would, but that's so DevOps.
22:12
@MikeyB because systemd attempts to create some cgroups prior to switching users.
So that wont work.
I think I can work around it by explicitly starting it with systemctl start user@<uid>.service somewhere.
pedal time
But, that makes it necessary for me to get involved ;) in something which in F20 is automatic/.
@MatthewIfe Just shoot all the users so that they don't try to run services that you have to support...
@freiheit psDoom?
22:17
posted on June 06, 2014 by Matt Simmons

This blog's web server has been up and down a lot in the past few days. The performance has never been amazing, but it's always been at least passable (unless I'm getting slammed by traffic, of course). But I finally gave up and dug into what was causing the habitual slowness. Running iostat on the […]

22:41
@MichaelHampton I found my problem.
It was deliberately removed.
I found the section of code used to spawn the service in f20. It isnt present in the systemd build in rhel7
When I searched for the line that was removed, I found this..
From the patches in the SRPM: 0200-remove-user-.service.patch
23:01
@MatthewIfe Deliberately removed, eh? That's an odd one. I wonder why.
Theres a bug reference.
Which, of course I'm not 'authorized to access'.
Like all of the god damned bugs you would ever be interested in working out the justification of.
systemd 208 is really quite buggy. There are a LOT of bugs that have been fixed in the last few months.
Well, theres been some issues with 208 session management that have been worked around in f20
I imagine it more stable in later releases i imagine.
@MatthewIfe Maybe RH will pull in some of the bug fixes. I certainly hope they do.
Pretty sure they will, with something as critical as systemd.
23:12
Um, now what?
Wait, what do you mean I'm not authorized to join the domain?
Switch to a vt and isolate to multi-user.target and back to graphical.
Oh, it's because ipa-client and sssd aren't installed. Stupid RHEL
> ORA-00932: inconsistent datatypes: expected NCHAR got NCHAR
Well, I was able to join the domain from the command line (and realm join mdmarra.brazzers.local is really nice) but the GUI is still b0rked
23:40
Or maybe I wasn't... it THINKS it's on the domain, but it sure as hell can't see any domain users or groups.

« first day (1364 days earlier)      last day (3903 days later) »