« first day (1150 days earlier)      last day (4117 days later) » 

14:02
@RyanRies PHILLY. MOVE HERE.
National Grid - I'd love if your site would stop apologizing for not working - and just fucking work.
Dicks.
@MDMarra I will, as soon I can get my GF to agree to it... she might as well be my wife... and she's picky about where she wants to live, which is funny considering that we live in the armpit of DFW Texas right now. Worst. Suburb. Ever.
I think Philly would be awesome.
Dan
Dan
@cole National Grid?
@Dan electric company.
Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck Fuck
2
Dan
Dan
14:05
@cole Hah, yes, I just Googled - it's British and I had no idea they had US ops
@RobM You just gave a lot of fucks
That doesn't happen too often I can tell you @RyanRies
@Dan do they suck as much there as they do here?
@RyanRies he gave them all away now, in case he had an urge to later.
I'm doing an audit/stocktake of my server rooms right now so I think I'm justified ;-)
Dan
Dan
@cole I don't think they do the same things
14:06
Ah
Well they suck ass here.
the server rooms do?
@RobM "yup, that's a server"
Dan
Dan
@tombull89 I think you can just asset tag the whole room as "Cloud"
@pauska What's going on with your VMware?
Dan
Dan
Yep, that's our cloud room. We don't know what's in there, because y'know - DevOps
14:17
@ewwhite I actually need to implement resource pools, due to a very power-hungry SQL server
@pauska Interesting. What's the situation?
Cluster size, RAM and VM profile?
@ewwhite Like I said to dan, we have different hardware in the servers.. same cluster is being used by developers, testers, lab etc
@RyanRies Except for all the hipsters.
@pauska What's different about the machines?
And this is vSphere 4 or 5?
and if the host with the most RAM goes down I need to be able to make 100% sure that our SQL servers doesn't get swapped or ballooned
@ewwhite read the backlog here so I don't have to type it out twice
vsphere 5.1
14:19
@pauska There's too much fucking in here... Hard to understand.
@ewwhite No there's not.. it's literally one page scroll up
144 and 256... how many in cluster? 3?
And what has the SQL server been allocated?
4 hosts, 3x144 and 1x256
@RobM You just doing inventory or are you actually assessing the rooms as a whole?
@pauska Enterprise Plus?
14:22
It's going to be allocated 96GB of RAM, which brings us up to 479GB of allocation
@RyanRies It is
I mean, it's no NYC, but that's what I like about it
if the 256GB host goes down we're overallocated (160GB per host)
@ewwhite Enterprise
I know you guys tried, but the guy is going down a spiral hard...
-5
Q: Is possible to install VMware ESXi Server on a HP ProLiant DL360 G3?

DarioIs possible to install VMware ESXi Server on a HP ProLiant DL360 G3 ?

so I just have to set up different resource pools to make sure that lab-stuff doesn't get powered on if the 256GB one goes down
@pauska If the SQL server is the outlier, you can do two things. You can reserve or your can set priority. Since this is one key VM, I would just set custom share values at the VM level.
And you can check this by viewing the worst-case allocation.
14:25
@TheCleaner It's like someone had a question in mind 8 years ago, wrote it down, and then found it today + posted it..
Windows 2003 and a G3 server..
I wouldn't do resource pools in this case, unless you're prepared to do the math to keep everything balanced as the composition of the lab environment changes.
Dan
Dan
@pauska :D
@ewwhite Yeah, resource pools requires you to recalculate the shares.. doing it per-VM is tedious as well
@pauska But you only have one outlier that you wish to protect.
Set it to a high priority, and that's your protection.
@ewwhite No, we have tons of production VM's that needs to be running in case of failure
the SQL server is just the reason why I have to implement this now
14:26
They won't not-run.
sorry, they need to be running in a fast manner :P
Have you experienced ballooning yet?
terminal servers in a memory blob is a nightmare.. etc
no, but I will if the 256gig host goes down
And are you running with any admission control?
160GB configured on 144GB physical won't necessarily balloon.
yeah
You're right, it could be shared.. could
14:28
@pauska But your VMs aren't using ALL of their assigned RAM all the time.
also true
hmm..
Especially in a windows environment.
so your suggestion is to just set shares on the most critical ones? would you set a reservation as well?
So I described this in detail to the CGI healthcare place here when they were in that 4:1 overcommit situation.
@pauska is the VM setup in a flat layout now?
no resource pools.
not right now, since I've been tinkering with it for a few hours
14:31
I'd go flat and look at the cluster "resource allocation" view.
Would you set reservations on VM's you don't want ballooned?
oh god.. are you running a single TS monster? bad, baad..
@pauska No... I only do reservations on VMs that either are running a heavy java application or where I need to guarantee a minimum about of RAM to the VM.
@MDMarra is supposed to expand it with a gateway sometime soon
@ewwhite heavy java application == sql server
same deal, it chokes itself with any type of ballooning
I'll take "Technology Tabloids for $400 Alex.
The answer: "Getting ready to run VMPlayer on a hosted VM." {ding} Cleaner?
TheCleaner: What is "things you do when you are desperate in IT?"
@pauska Java apps like Websphere and Tomcat... where VMware doesn't have insight into the memory allocation inside the VM.
14:34
@ewwhite or SQL
@TheCleaner VirtualBox
@ewwhite A broker, and we're doing that for Valley?
@pauska Go flat and just start by setting the shares of the SQL server to high. Shares are evaluated on the sibling level, so it needs to be all VMs or all resource pools, but not a mix.
@MDMarra I still need them to agree not to try to create file shares on their own
haha
your customers are the worst
run rds vm's with no more than 4 vcpu's and 30-35 users each is what I recommend
scale the memory according to what software they use
14:37
@pauska I'd love to... but there are probably 90 users on now.
3 vm's then
and a mismatched VM host situation just like yours.
"We don't have an IT staff. Do you think it's OK if we have the payroll manager run Exchange, AD, RDS, and file services?"
Dan
Dan
Seems legit.
I've seen setups where people set up lots of ts'es with 1 or 2 vcpu's.. don't do it, it's stupid
the internal fair share cpu scheduler for RDS beats the hypervisor scheduler by a mile
Dan
Dan
14:38
@pauska This is a throwback from where more was a way to get round RAM limitations in 2003 and, to some degree, 2008 x86
@pauska I have two different customers with the same project going on now (SCSM and SCORCH implementation). One is going to virtualize everything (including 16 vCPU database servers) and the other one won't virtualize anything with more than 2 vCPUs.
at least for windows RDS - I have no idea if Citrix has some vodoo tricks to make the hypervisor understand RDS scheduling
@Dan yep
Dan
Dan
@pauska Nope, basically the same
we had 15 2003 vm's with 2 vcpu..
@MDMarra payroll manager's desktop FTFY
14:39
@TheCleaner Yeah, we may have had ESXi consume NFS storage provided by a VM on the same host for a little while. We were temporarily desparate for space.
Dan
Dan
@pauska Yep, that's the job! I actually set something up similar.......last year!
@pauska Setting the shares alone will take care of your server. But if you want to do some cool stuff with custom HA slot sizing, that could be fun, too...
Dan
Dan
Virtual Server 2003 TS servers on brand new G8 kit :D
@Dan wat
@Dan A greenfield Windows 2003 setup? Are they aware of April 8th?
@MDMarra 16 vCpu, holy moly..
14:40
@pauska Logicworks did a greenfield 2003 installation of a few hundred VMs over the summer.
but the "no more than 2vcpu" is outdated.. the modern hypervisors doesn't lock that way anymore
@ewwhite :(
and 2003 x64 database servers on DL580 G7's...
and using Isilon CIFS as a fileshare for all of them...
@ScottPack an inventory. What is in what rack in what room, should it be there, should it be warrantied, is is warrantied, is it active/spare/? , should it be scrapped, etc.
@pauska Data Warehouse, bro
I'm happy you left that place
14:42
SSAS <3 CPU
what fun
Dan
Dan
@pauska It was done under duress - it was a quasi P2V (I actually did a rebuilt, but to the same spec) due to licensing
and having the Isilon's BSD stack crash the CIFS and lose much data...
@pauska It's opened up so many doors.
Dan
Dan
@pauska I'm under the impression it's somethign to very aware of. I'm sure vSphere still waits till [x] cores are free
2003 server? I don't even know how to prime the steam boilers on those any more.
14:43
@RobM Necessary work, but still drudgery.
@Dan not as much anymore. VMware has gotten smarter.
@Dan Nah, the vSphere CPU scheduler was "fixed" in the 3.x line to not lock until all threads can be processed
And Hyper-V never did that, to the best of my knowledge
It certainly doesn't now.
You may want to consider NUMA and other things like your server platform...
Dan
Dan
@MDMarra Really? It's still mentioned in that vSphere 5 book
Which book? :)
The one from Lowe?
Dan
Dan
14:44
@RobM Biggest thing I noticed was that, in 2008 onwards, you can type a command into any explorer window and it'll run, or just "cmd" to get a command window. You can't do that in 2003 and it starts a bloody IE window and tries to search
@MDMarra Yeah, at least I think it does
I gave my copy away
Great book, but I don't need it anymore
Dan
Dan
Yeah, it's in my car. I so wanted to do my VCP, but just can't justify the costs and work won't pay :(
@MikeyB This one is more of a "we have plenty of storage assigned to this VM and it has plenty of hardware resources assigned...so instead of paying for another VM, we are going to host a VM within a VM" :)
@Dan Is your work a VMware partner?
@TheCleaner facepalm
Dan
Dan
14:46
@MDMarra I'm not sure what our current status is actually - we were once, at least. Immaterial now anyway - if they pay for anything I'll wind up being liable to pay em back when I jump
@dan its small stuff like that which makes a big difference
@TheCleaner That's weird.
@Dan Right, but there's a promo code for partners now where you can skip the expensive class
So you could get a VCP for the $225 exam only
which isn't much to pay back
@MikeyB welcome to my world. Doing things the "right" way is simply a state of mind. Results=right.
Dan
Dan
@MDMarra Oh right - cool. No, it's not - don't think I'm good enough though and not sure I have the time to really put in the work. I may take a look when I'm in the office, though
It's been about a year since I got that book, actually. It was a great read, though
14:47
If your VMware myLearn account is registered with a VMware partner, code ACCELERATE lets you skip the class
Dan
Dan
Cheers
Although, who knows if it works outside of the US
@TheCleaner Fortunately, most of the time I get to live in "Maintainable results == right"
@MDMarra You be hustlin'
Yep
@RyanRies you know AD CS right?
14:51
Yes
We have a customer rolling out Cisco ISE for BYOD and user/device based cert-based .1x auth for wifi
They called up after the project started to assess their PKI
They have an enterprise root colocated on a file server
only issuing LDAPS certs now
14:52
Oh god
And were shocked when I said it would take 5 days to plan, design, and implement proper PKI
I mean, it has Infrastructure in its name...
If anything, I thought I undershot the estimate
"Oops"
They said "can we make it work with what we have now and change it to best practice later"

"sure, if you revoke every cert and reconfigure ISE and anything else to use the new infra"
14:53
Yeah nothing pisses me off more than half-assing and band-aiding shit together to try to get something done for cheaply, and get it done yesterday, and that will solve all your problems
AD CS is a royal pain to get working correctly...
The good thing is that if we do have to do stuff like that, we dont have to maintain it after
But once it is working, it just works.
Though we do always try and push them in the right direction
14:55
Is there a defentive list for RAM/board compatability for HP Machines? One site I'm looking at says it does support the RAM I have and another site says it doesn't.
@RyanRies so, long story short, I'm the only one on staff with enough PKI knowledge to do this and I don't even have a ton of AD CS experience...and I'm booked into March
so.....customer is SOL right now, haha
In other (old) news, blackberry singing its death song
@tombull89 Servers?
@ewwhite Yes
@tombull89 Yes. What are you lookin' for?
14:56
@MDMarra Sounds like just the kind of thing that I could help with if I lived in Philly!
We're considering bringing someone in on a 5 day contract to do it :)
Though we'd HIRE YOU DIRECTLY TO A SENIOR ENGINEER POSITION IF YOU'D MOVE HERE
@ewwhite I have 499277-061 as my RAM modules. I'd like to know if it's supported in DL385 G5.
@tombull89 UK?
and is it a G5p or just G5?
Hi
@Dan You should be careful and not just create every VM with 4 vcpu's, cause that could lead to severe locking (all the servers start their antivirus scan at the same time etc)
15:01
JVM sending SIGSEGV in random instructions and lots of zombies accumulating. Could that be a failing memory stick ?
But it doesn't need to lock 4 cpu's when a VM with 4 vcpu's needs a single thread.. that was fixed back in 3.x somewhere
@ewwhite Uk and G5.
Dan
Dan
@pauska I tend to give most "normal" servers 2vcpus by default, and assess anything else as it comes
@tombull89 G5 uses PC2-5300 DIMMs
Dan
Dan
@pauska Are we 100% on this
15:03
@Dan it's fine these days
@ewwhite shakes fist at SS
Dan
Dan
Cool. And here's the in depth info:

http://www.vmware.com/files/pdf/techpaper/VMware-vSphere-CPU-Sched-Perf.pdf
Any way to find out what does support it?
@Kwaio Because java.
3 mins ago, by ewwhite
@tombull89 NO. See page #39 - http://h18000.www1.hp.com/products/quickspecs/13012_div/13012_div.PDF
15:04
@NathanC i'm not sure zombies are related. But my tomcat doesn't start anymore, and that's production \o/
Otherwise, I would agree. I hate java...
@Kwaio Are you running on proper ECC memory on a server that'll throw an MCE on a memory failure?
@MikeyB I'm not sure. that is a G7 running ESXI 5 with a single RHEL3 VM
I suppose HP pro servers have ECC memory
yes, they do
you can use unregistered ram in HP servers
it's a violently stupid idea, but you can
@Dan well unless someone tells me otherwise, I'm sure.. I did a lot of research on it back in the days where the fix finally came
@Kwaio Likely not your memory then. Probably RHEL3: h30499.www3.hp.com/t5/Systems-Management-OpenView-OP/…
15:09
i have dozens of others identical system that run fine =/
@Kwaio run the memory diagnostics on it
Dan
Dan
@pauska Seems I'm wrong indeed - perhaps the things I've read have just overstated the current status
@Kwaio Well, support for RHEL3 ended 3 years ago, perhaps it's bit rot that's starting to accumulate in the memory buffers?
@Kwaio if you have the HP tools installed in RH, run hplog -v
I thought they can't be installed in a VM ?
15:11
It's a VM...
oh I didn't know this was virtualized
vSphere console may show you RAM status
ESXi? licensed?
@ewwhite yeah, with a proper license
DL380G7 => ESXi 4 or 5, not sure => RHEL3
(Because RHEL3 isn't compatible with G7 hardware xD)
I'm allowed to laugh because i would never have done that...
But for the time being i'm in charge of supporting those...
That's ESXi 4.1
I'm confused...user is having weird permissions problems.
We have user folders for everyone ...they have "full control" over this folder and all subfolders. But, when someone creates say User Folder -> Folder -> Folder, they don't have access to that 3rd level folder...permissions don't carry
but everything's supposed to be inherited...
15:19
@NathanC OS X?
windows
fire. And lots of it.
@NathanC Have you checked to make sure that it's inherited on that third level folder?
I know that's my solution for everything but it works...
And do your users have "modify" or "full control"?
15:21
@MDMarra Yes, it's checked...but it's only carrying over a couple users
as in, the administrator and one of the domain admins
full control
Yeah dont do that. Do modify
Hm, now the permissions are filtering through...does full control not inherit or something?
it does
it should
before what would happen is a subfolder would have "special" privs that said "applies to folder only"
15:23
sometimes weird shit happens if its been forked with
and it was inherited by the user's folder
So there are a few things that are special about NTFS that could cause this
but, anything under that lacked the permission...so user could create stuff but couldn't do anything with it. most bizarre thing heh
A folder that is "moved" i.e. cut and pasted on pre-2008 will retain its original permissions and not inherit
@MDMarra User right-clicked -> New -> Folder
15:25
@NathanC over a share or locally?
Full control allows users to set permission and owner. This sucks because sometimes applications set owner and permissions and if the user has full control it will work
Give them modify, no need for full control 99% of the time
@MikeyB Over the share. It's also DFS if that matters.
If they literally just went "right click, new folder" and it didn't inherit properly, then theres something fucked up in your upper-level permissions somewhere
@MDMarra Ah, fair enough. I have to audit our "Users" folder since permissions are wonky
some users have "Everyone" set to readonly but not themselves as modify...when I started "Everyone" was being inherited at the root folder so anyone could view anyones folder if they knew how to get to it
i was like ..."nopenopenope"
@NathanC does mkdir behave?
15:31
@MikeyB I can't test myself since I have administrative rights...so it'd work regardless :P
I feel like the "CREATOR OWNER" account missing from some folders might have something to do with it..
@NathanC check that the permissions you want it to inherit isn't set to "This folder only" or something
and like Mark said - never give full control to anything at all unless you're 100% sure it requires it
exceptions for the SYSTEM and local administrator group of course
Hey there
sorry guys but anyone here knows some cheap and functional ssl certificate to buy?
or free at all is better
we use startssl SSL but damn it doesn't works on Firefox :O
free translates to bad. always.
If money is all you care about, get one from GoDaddy
it works on everything these days
WTf is wrong with you java ?
i mean except the fact that you EXIST
@pauska thanks
i'll check godaddy one
does that shows green https part of the browser url address bar?
Dan
Dan
15:43
@sbaaaang Yes, because it'll be trusted by a known Root CA
Make sure you get it for the exact domain you need (including the subdomains)
If you want to use it on multiple subdomains, you'll need a wildcard
thanks dude!
posted on November 04, 2013 by Mark Marra

Full control is a set of permissions that I see granted quite, perhaps more frequently than it needs to be. For example, I see quite a large number of customers with the full control NTFS permission for each user set on their network home folders, or a group will full control to their departmental network shares. I'm a firm believer of using modify permission inste

@NathanC in most cases Creator Owner isn't required for anything. The exception is if you're doing stuff like "Subfolders only" permission to "write/append data" or whatever it is to automatically create redirected folders with least permission
In most cases, you don't need Creator Owner and in fact, shouldn't have it unless you know what you're doing.
Ah...okay.
I have a feeling I'll be going through each user's folder and reassigning permissions...
Powershell, bro
15:51
well, 90% of the users don't even use the shared drive which is interesting
and powershell it shall be :p
let's try sun jdkl instead of openjdk
@tombull89 I got an error there, but I heard about those
I can't wait to see the 20TB HAMR disks they predict in a few years
Customer: "We need you to get us 50x400GB 1.8" SSDs in 2 days."
Us: *What? The fuck you think we have in stock? You should be ordering that shit WEEKS in advance!* "Sigh... we'll see what we can do."
Tada ! upgrading the jdk doesn't work
16:07
@MikeyB Whackcherry?
Every obnoxious, large-ish scale request you get, I immediately assume it's them.
@Wesley Lickmynuts 'arry
@MikeyB Not in front of company, dear.
@Wesley Bad time anyways, sac like a blood orange.
Evidently my internal censor is offline today.
@MikeyB Man you're like a walking genetic Fukishima.
What else could cause the JVM to SegFault ? It's obviously not the memory, it's not the JDK...
16:09
Yeah you might have had some negative external influences, but your design was flawed from the beginning. =P
i have free memory, free disk free inodes free cpu time
@Kwaio The...application?
> This is NOT a place for 'Live Support'
But you know I love you anyway @MikeyB
@Kwaio just blow away that VM and clone one of the working ones.
@Wesley D'ahh..
16:10
@MDMarra Probably
@Iain Apologies.
Just how big is @DanBig?
@MikeyB Can't. We have 1 server per store running ESX with a single VM taking all resources
I'll try to rollback the backups to a ferw days ago
@Kwaio Oh, you're only using ESX as a HAL?
B=============D
4
@MikeyB RHEL3 is not compatible with HP DL380 G7 hardware =P
@MikeyB I know how ridiculous that sounds
16:12
@Wesley We happen to have 16 of those drives in the back from my storage testing... sure... whelp time to say "LOLNOPE".
How is everyone else's close vote trigger finger doing?
FIRE ZE MISSILEZ!
@Kwaio yeah you're solving the wrong problem. Move off RHEL3.
uh...yeah...
16:13
@MikeyB Wow, but still - 50? Sheesh what are they doing? Acting as a CDN for @ewwhite's porn sites?
room topic changed to The Comms Room: This is *NOT a place for 'Live Support', ask on the main site. // 4 November, 1960 – At the Kasakela Chimpanzee Community in Tanzania, Dr. Jane Goodall observed a chimpanzee using a grass stalk to extract termites from a termite hill, the first recorded case of tool use by animals. There's a joke about users in there somewhere. [be-the-change] [pwned-by-a-wild-markdown]*
@DanBig Haven't been on the main site. Don't have enough cough medicine to withstand it.
@MDMarra You forgot "Does not apply to Ed White"
The first part's always been there
only 1960?
According to wikipedia's "on this day" section on the front page
16:16
@MikeyB I have no power to do that unfortunately. and i'm gettiong out of here ASAP... which is a couple month away
Yesterday I did a bunch of yard work. Namely, I cut down a bunch of branches that were hanging low enough to scratch my car. While doing so, I sustained a fair number of small lacerations... Including on my face! That'll look great in the interview today! ;)
@KevinSoviero "Yeah, about that, I got attacked by a Redneck Tree on the way here."
@pauska your EOS date for 2003 is off by a year
XP dies April 8 2014, 2003 R2 w/ SP2 gets to live until July 14, 2015 - support.microsoft.com/lifecycle/search/…
@MikeyB Kind of wish I hadn't read that...
16:28
This is some MacGuyver shit...
3
Q: Is it possible to boot from a passthrough LSI 2008 SAS controller under VMware ESXi?

KurtI currently have Solaris 11 Express running on bare metal. I'd like to convert the machine to an ESXI host and run the existing os as a virtual machine. I have ESXI setup, the virtual machine made, if I boot the vm from the recovery disk the drives and lsi2008 controller is visible. Everything l...

Pretty sure I violated space-time causality the other day. I mucked with the phone-system 2 hours after it suddenly started acting up.
@ewwhite LOL... as it turns out this is EXACTLY what I want to do at home. I have an LSI2008 IT controller with drives attached and I want to boot FreeBSD to do ZFS in a VM with the LSI card attached to the VM.
@MikeyB It works. I do it. But you need something to hold the VM.
I payed the hosting for 5 years, thats why i dont change. Nathan C can you stop voting the question negatively? — AlexSim 43 mins ago
I love when people assume
but, the question is crap ._.
@MikeyB Do you also want to have other VMs reside on the storage shared by the FreeBSD VM?
(I've done this quite a few times... in the end, I tear it down because it's like sucking your own d*&k....)
16:43
@ewwhite potentially, but probably not. I'd be moving it to my libvirt server so I could do something like ensure the storage pool is started after the VM is started. In reality, probably not, I'd probably just keep using it for media and backup, but I'd get rid of a computer.
ZFS runs fine in a VM... even without a passthrough controller.
I hate it when the software we use in production requires a GUI screen for command-line operations... -_-
So, we have to run an import task in an open Windows session on the server so it'll work. What is this, 2001?
I think one of the best parts of my day is seeing what's on the star wall in the morning.
@84104 Usually d*cks, barrels of lube, or some random out-of-context starspam
Only on The Comms Room.
@NathanC It's the little things, isn't it? =P
16:56
Anyone else notice the little war being played out on the star wall?
Some people are trying to imbalance the names, some are trying to balance them
This will end well.
Some sort of dysfunctional popularity contest.
Yeah if somebody could unstar that 8-star line, that would be great.

« first day (1150 days earlier)      last day (4117 days later) »