Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Locked thread
Scikar
Nov 20, 2005

5? Seriously?

It's a bad sign when they ask about your previous salary and there isn't really an easy way past it if you actually want to work there. They have no more right to ask this than you have to ask how much they are getting paid, and it will only be used to lowball you either now or later on, so don't tell them. Lying is dangerous because you never want to be in a situation where your manager doesn't trust you, and you don't want to be caught out when you later chat with a co-worker you make friends with about how badly you were paid in your last job.

Better to say that unfortunately you can't discuss your previous salary as it is covered by NDA (history with a legal firm probably helps plausibility here!), even if that's not strictly true. Firstly you're much less likely to accidentally give this away, and secondly you can just write yourself an NDA that says "I promise myself never to reveal my salary history to another employer", sign it, and you won't be lying.

Adbot
ADBOT LOVES YOU

Scikar
Nov 20, 2005

5? Seriously?

skipdogg posted:

Hahahaha Nah. When people would leave the company I'd dump their email to a PST, and you just kinda poke around

Is this really a thing in other IT departments? Like any sensible person I treat my emails as though somebody I don't want is probably going to read them one day, but that doesn't mean I'm going to be that person for one second.

Leavers' mailboxes get archived, nobody touches them without HR approval. On HR's word I will give someone else access, but I'm not looking at any of it myself if I can avoid it. On the rare occasion that it makes sense for me to search someone's mail or the journal, I get a ticket with precise details of what I'm looking for and email approval from my manager at absolute minimum. Ideally I'll have HR approval and get my manager and the requesting user to sit through the search while we do it but I'll settle for just my manager's approval if I have to since I can trust him to take responsibility for it. Regardless of that, I'm certainly not going anywhere near other people's emails just out of pure curiosity. That's a dark road that leads to bad places.

Scikar
Nov 20, 2005

5? Seriously?

Did Postini detect the virus or did it just classify it as spam?

Scikar
Nov 20, 2005

5? Seriously?

Swink posted:

Question: Those of you who map drives to a DFS share, have you ever had any trouble with file conflicts? I have two designers who work on the same files out of a single DFS share and I kind of feel that its only a matter of time before they open the same file at the same time and cause a conflict.

DFS is simple last-write-wins and it has no mechanism for dealing with a file being edited in two different locations. We use it with multiple targets enabled for home folders where only 1 person is going to be accessing the files, but any shares where more than one person accesses the data should only have 1 target enabled unless you like pain.

Scikar
Nov 20, 2005

5? Seriously?

Che Delilas posted:

What the gently caress is a couple thousand dollars paid out to people to make sure they will be willing to bust their rear end to get poo poo back online?

A couple of thousand more than they think they can avoid paying and add to their own bonus instead.

The stupid thing is any decent staff are going to care enough about their systems to make themselves reasonably available anyway. I do 1 week in 5 of 24 hour on call, for critical systems only (and not directly with end users ever) for which I get a monthly payment, plus overtime for any actual work I do and I can take that same amount of time off the next day. While on call I have to respond to a call within 30 minutes (so I don't even have to actually answer the phone) and I'm expected to keep an eye on my email for monitoring alerts which is just glancing at it once an hour or so while I'm awake.

Since this is a perfectly reasonable policy, everyone on my team goes a bit further to help out, so if something comes up I have no problem with whoever is on call dropping me an email to ask for some help/advice and if I spot a minor issue while idly checking my phone I'll sometimes just jump on and fix it since I'll have to do it later anyway. But if that standard policy was ever bumped down you can guarantee that outside of work hours my phone would be off unless I was getting overtime.

At the end of the day it's just good management, treating your employees with respect and reasonable remuneration isn't even expensive, but if you do that they'll move heaven and earth for you when poo poo goes bad.

Scikar
Nov 20, 2005

5? Seriously?

Caged posted:

I believe there's still quirks in Windows about picking up an admin token, in that even if you're logged in with an admin account you still need to elevate.

That's not really a quirk, it's kind of the whole point.

Scikar
Nov 20, 2005

5? Seriously?

Lord Dudeguy posted:

So how the heck do you do it? Do you basically have to pre-emptively robocopy the entire volume again before re-introducing the node? If that's the case, then I'd need to set DFSR to manual service startup and not automatic, which is the default.

:edit: ^^^ Dear christ. I'm going to start shopping for a better method. poo poo, SAMBA w/ rsync has better failure recovery features than that.

You don't robocopy it. As soon as you remove the server from the replication group and the others in the group pick up the changes from AD, they will refuse to replicate with the problem server whatever happens. When you add the server back in it is in read-only state until it is caught up with whichever server you picked as authoritative when you added it. DFSR itself is usually way quicker than robocopy unless you have a short fat pipe between the two servers, and even if you pre-seed the data it still has to process all of the files to check they match and build the DB anyway. In 2012 R2 you can export the DB from another server and import it to a new one if you're cloning them side-by-side as well but I haven't had chance to do that yet.

Scikar
Nov 20, 2005

5? Seriously?


Even in 2003, if memory serves you can still extend partitions when it's running, just not the system partition.

Scikar
Nov 20, 2005

5? Seriously?

evol262 posted:

Ah, the true goon, critiquing people in the most puerile way possible for the subjunctive mood and proper homophone usage which are so crucial to writing Java.

I'm pretty sure he wasn't being completely serious, but at the same time I would have thought that not being able to write correctly in your native language would cause some problems writing in a programming language. And even where it doesn't, it certainly could lead to that person writing documentation that is ambiguous or inaccurate. I work with several people who are extremely competent technically but cannot compose an email or write up project documentation without making spelling and grammar mistakes throughout, and I am 100% certain it holds them back career-wise.

Scikar
Nov 20, 2005

5? Seriously?

Daylen Drazzi posted:

Who the hell knew that having the LAN set to SHARED instead of DEDICATED would prevent the fucker from communicating?

I would guess you probably got to the bottom of this (or this is sarcasm), but just in case, if these work like IBM RSAs then dedicated means "use the specific NIC for the DRAC", and shared means "piggyback off one of the regular NICs for the server itself". So depending on your network setup that might explain why you couldn't see it. Handy if you're limited on switch ports or are shipping the server off somewhere and you know that an idiot will be plugging it in at the other end.

Scikar
Nov 20, 2005

5? Seriously?

Lord Dudeguy posted:

- "The phone beeps 4 times before ringing" = The phone beeps 4 times before ringing. (gently caress you, Lync certified phones, not being fast enough to negotiate codecs)

This might sound weird but check the handle count for the services on your edge server. We had this exact issue but we noticed it went away if we rebooted the edge, and on closer look when it happened the service would completely lock up if you tried to restart it and there was an obvious handle leak in the service executable. The only fix was to reboot the box entirely. We reported it to MS but they wouldn't look at it until we had everything patched, which we couldn't do due to issues with our combined 2010/2013 environment. It went away at some point but I'm not sure what the permanent fix was.

Scikar
Nov 20, 2005

5? Seriously?

Dick Trauma posted:

EDIT: Goddammit this is still too long and makes me sound like a colossal whiner.

Not really, more like colossally unlucky. You're right that it's lovely management, competent managers know that people don't just magically become poo poo overnight. If their performance was excellent but has dropped it's far safer - and less work - to figure out the problem and resolve it than it is to get rid of them and try to find a good replacement. Screaming at them and doing nothing else is guaranteed to just make the problem worse. I'd like to think I'd come out fighting in that situation but honestly I'd probably react the same out of pure shock, I'm stunned that anyone can actually think ambushing your staff like that is ever a good idea.

Scikar
Nov 20, 2005

5? Seriously?

Lord Dudeguy posted:

Oh my god...

The team was mentioning that during an edge outage the phones were dialing great. I didn't believe them.

I'll try bouncing the edge tomorrow. Why in the hell would it do that? I have federation disabled.

:edit: 6,100 handles for System process. Restarted Lync services and it dropped to 5,100 handles. We'll see what happens tomorrow.

It's definitely a weird one, we couldn't believe it was the edge either. I spoke to the lync guy just now and he says the root cause was DNS, we had old entries from our 2010 environment that weren't updated for the new 2013 servers. Instead of failing gracefully it causes the system handle leak on the edge and leads to the dreaded beeps. He spotted the incorrect entries while looking at something else in DNS, updated them and we haven't had a problem since.

Scikar
Nov 20, 2005

5? Seriously?

Lord Dudeguy posted:

gently caress

After a knock down, drag out :sotw: fight between my boss and I, where I insisted the Edge services were working OK and he begrudgingly started calling Polycom for a possible refund...

... it was the Edge. There were no static routes for our VoIP VLANs on the Edge LAN interface. Wasted a Software Assurance ticket with MSFT to tell me that.

loving hate humble pie. I like the peg that I'm on. gently caress getting taken down a few.

Well, I'm glad it got fixed in the end at least. :shobon:

Out of interest, what made you put VoIP on its own VLAN? One of the things our networking guy liked most about going for Lync with the better Polycom handsets was the fact we didn't need QoS any more with RTAudio, so the phones don't need to be on a separate VLAN. There's one site with a 1Mbps connection but calls on that site are rate limited with CAC to 128kbps total and that's all we've needed.

Scikar
Nov 20, 2005

5? Seriously?

DrAlexanderTobacco posted:

The first thing you should do is always ask for more info. What are you trying to do? Is it slow to start up as well? How long does it take to open internet explorer? What about files from a network drive? poo poo like that.

Resmon.exe (In windows by default but not advertised that much) is great as it'll break down resource usage in great detail; more-so than task manager.

It might be slow simply because it's a poo poo PC. If it's got a low (2GB) amount of RAM, or a 5400/7200RPM disk that might be the bottleneck. Try to remember that when someone calls in with :frogsiren: THE COMPUTER IS SLOW :frogsiren: that doesn't mean it's suddenly happened out of nowhere. They could just be wise to the ways of the helpdesk you're at and know that screaming about something will warrant a closer look. Whereas in fact, the laptop is 3 or 4 years old and they've put up with it until now.

Yeah Resource Monitor is really the way to go for figuring out why things are slow. On anything without an SSD, chances are pretty high that it's going to be disk related and resmon will give you a breakdown of which process have the highest disk activity. But always get more specifics from the user first, otherwise you'll spend ages troubleshooting just to find out that they were trying to watch youtube on a 2 meg shared connection.

Scikar
Nov 20, 2005

5? Seriously?

Pyroclastic posted:

I'm just a lowly computer janitor. I can't go lugging servers around (and this one may be virtualized now), especially since the other two techs are imaging off it in their own buildings.

Maybe if our bench tech had stayed on he could've learned how to make SCCM more flexible like that, but I figure we'll be damned lucky if it even stays functional for two more months. Our SCCM server crashed twice last year and each time it took over a month to rebuild because the backups didn't work (one admin said it was because of file pointers and symlinks or something which don't get backed up correctly, so the entire deployment service had to be rebuilt).

Our old wiring generally splits the 4 wire pairs into 2 separate jacks, but I'm fairly sure I reterminated that cable with a new module that's actually 100mbit. Otherwise, I probably would've been lucky to finish 3 computers in a day.

SCCM doesn't have to have everything on one box. The actual packages and image files are taken from a distribution point which doesn't have to be anything special. So you can set up a laptop as a distribution point, sync all of the stuff you need for deployment to it, then move it to the site containing the subnet your lab uses. Then you just plug the laptop in in the lab and the machines there will pull the data from it. This can cause some problems if someone updates a package on the laptop DP while it's off or on a slow site though.

Scikar
Nov 20, 2005

5? Seriously?

Crowley posted:

I have a ticket to get a list of department membership from a list of usernames. Luckily for me the users' OU is their department. Unluckily for me I apparently can't Powershell for poo poo.

Can someone please tell me what I'm doing wrong in the following script?
code:
### Define Variables before we start ###
$PathToFile="C:\path\etc\blah\input.xml"

### Open XML Document ###
$xdoc = new-object System.Xml.XmlDocument
$file = resolve-path("$PathToFile")
$xdoc.load($file)

### Get the field we want from contents ###
$BrugerID = $xdoc.SelectNodes("//BRUGERID")

### Get the AD OU from the UserID ###
get-ADOrganizationalUnit -LDAPFilter '(name=$BrugerID)'
Everything works until the last line. If I Write-Output $BrugerID I get
code:
#text
-----
DOMAIN\userinitials1
DOMAIN\userinitials2
which should work fine, but I still get a big empty nothing out of get-ADOrganizationalUnit.

I've been wondering if I'm using the $Variable wrong, should I use Where-Object instead, or.. something else?

Get-ADOrganizationUnit means you're trying to find an OU that matches some properties. A user isn't a property of an OU, it's an object within an OU. But the directory is a hierarchy, so the parent OU of an object is a property of the object itself.

If you do

code:
Get-ADUser -Identity $BrugerID


you'll get the actual user. Pipe that to Get-Member:

code:
Get-ADUser -Identity $BrugerID | Get-Member
and you'll get a list of the basic properties of a user. Get-ADUser is actually limited by default so it runs quicker, but you can do:

code:
Get-ADUser -Identity $BrugerID -Properties * | Get-Member
and you'll have an enormous list. I'm not entirely sure if you can straight up get the OU (and I don't have access to a domain right now) which would be a quick way to get the OU itself, but one thing you always get with an AD object is its Distinguished Name, which is its full 'path' name in the directory. So:

code:
$user = Get-ADUser -Identity $BrugerID
$user.DistinguishedName
will give you something like "CN=Glen John,OU=NorthAmerica,OU=Sales,OU=UserAccounts,DC=FABRIKAM,DC=COM". You can use the latter part of that as the -Identity in Get-ADOrganizationUnit to get the actual OU object if you need it, though trimming the start of a string to a given character isn't something I can do off the top of my head (you can trim to an index value so you find the index of the character you want and then trim to/from that value).

Scikar
Nov 20, 2005

5? Seriously?


The shadow behind the footprints is my favourite thing.

Scikar
Nov 20, 2005

5? Seriously?


Piece of cake!

https://www.youtube.com/watch?v=B7MIJP90biM

Scikar
Nov 20, 2005

5? Seriously?

odiv posted:

Jesus Christ, if you put a ticket in at 10am and then are unreachable for half an hour, and then when I finally get a hold of you, you complain so much that it takes ten minutes to fix your problem instead of five, don't loving bitch to me about how much work you didn't get done this morning and "most of yesterday" because of it.

I mean, I'm going to do gently caress all about your attitude except apologize because I abhor confrontation, but it's a lovely thing to do.

Don't ever apologise for that poo poo. Maybe "I'm sorry to hear that", but never offer a straight up apology for something that isn't your fault. You're just validating their unreasonable stance and making it worse next time.

Scikar
Nov 20, 2005

5? Seriously?

Dragyn posted:

This is fairly common practice in Health IT, they're not looking for you to prove it, they're looking for you to confirm it was deleted/destroyed and not retransmitted anywhere else. It's a CYA measure so that the person who sent the erroneous data can say "the breach was contained and eliminated".

Right, but it's probably still a mistake to respond with anything other than "where do I send the bill to for our time spent on this?". I'm not going to go through tape backups of mail server's archive to remove an email that shouldn't have been sent to my org in the first place, and I'm definitely not going to open myself up to getting sued by saying I've deleted it only to discover it was still on a laptop that had cached exchange mode running and then got stolen or something. I might be inclined to do that as a favour to the IT/security team of a company that mine has a good working relationship with, or even that we don't providing they ask nicely, but someone I've never spoken to sending legalese like that is going right to the back of my queue.

Scikar
Nov 20, 2005

5? Seriously?

fishmech posted:

Well the problem is that the ones the station requires are over a meter tall, and quite a bit more complicated than a normal valve/vacuum tube is.

This isn't one of the ones they'd need to replace, it's a similar design though, but smaller:

Apart from the specialist tubes for generating xrays and stuff, I would guess the only significant market left for standard valves is guitar amplifiers now. And as far as I know the only country still manufacturing them is Russia. There are some US brands but I think they're still Russian-sourced.

Scikar
Nov 20, 2005

5? Seriously?

Judge Schnoopy posted:

What's the best way to get an accurate look at memory consumption across three hosts in a VMWare environment?

I'm planning on increasing memory in a few guests that are below best practice specs (and are running like dogshit). My MSP is crying fowl that this will push me to 90% memory consumption according to the Host Summary page in VSphere if we lost a node, and that it could start to cause serious ballooning problems.

I'm looking at an ESXI host with 64 GB RAM. The sum of allocated memory for all guests on the host is 27 GB, and it's eating 2 GB overhead. On the Host summary page it's reporting 29 GB / 64 GB used.

There's no way all of the hosts on this server are running at 100% memory utilization. So how do I find out my actual utilization? Is this a safe practice to allocate 90% memory in a failsafe environment if the servers are utilizing less than 50% on average? Or does the host not care, committed memory is consumed memory, and 90% allocation is a dangerous game to pley?

In my research online I can't find a straight answer on this because I know over-allocation is a thing some people role the dice with, which leads me to believe that there is some metric of utilized guest memory that I'm missing and is important when calculating redundant capabilities.

The host knows that not all committed memory is actually in use (I think the metric in vSphere is %active or something like that). The rest hasn't been written to for a while so while it might still be getting read, it also could be freed memory that the OS doesn't have allocated any more. By default the host only tries to reclaim this with the balloon driver above something like 95% usage, so depending on the profile of your guest VMs you might have way more failover capacity than you think. The risk is that you don't want to ever have the host allocate swap when a VM requests memory, since performance takes a massive nosedive at that point. This is an actual problem, ballooning on the other hand has fairly low performance impact (it encourages guests to use their own virtual memory on disk, but the guest OS can be smart about this unlike the host).

All that said if you're in this position then you'll need to buy more RAM as soon as you need to spin up a new VM for some project anyway, so you might as well just buy it now.

Adbot
ADBOT LOVES YOU

Scikar
Nov 20, 2005

5? Seriously?

Dr. Arbitrary posted:

Next big idea (Don't Steal)
Some way to aggregate a bunch of CPUs into one gigantic Meta-Core, so that applications that charge per core just get stuck with one really powerful one.

Oracle are already one step ahead of you with that Core Factor bullshit.

  • Locked thread