@HackToHell why are you an explosion in the room list but a girl in the chat lines? is it some kind of weird bug where it doesn't update your mini-gravatar?
@OliverSalzburg changed his gravatar and it updated like the same day in chat
Could use some advice: I've been tasked with moving everyone over to our new Citrix XenApp servers. Think I'm okay to move all +350 over in the course of a couple days, or should I take it easy? I'm worried some problems will show up along the way. (There's 6 vm's on two physical machines running the farm.)
there's something called load testing and you do it in an environment called integration
what, you don't have integration? fires bullets at you
btw that's NARQ
that's just being a worry wart
never worry; just measure (and yo dawg everything you do)
as in, yo dawg i herd u wanted to move everyone over to your new Citrix XenApp servers, so we got u a Citrix XenApp server so yo can move while yo move
well, let me put it to you this way: people/companies/managers who say they don't have enough money or influence to procure a validation/integration environment should be prepared to work out any bugs that arise in "validuction", i.e. validation and production in the same environment, i.e. you hit your bugs in production and have to deal with the users complaining
well your question was a general worry about potential problems related to load, so if you don't test it, there's no way to know. you could get generic advice about how much the server should be able to handle, but it's all a Wild Arse Guess until you actually put the pedal to the metal
you COULD move people over in batches gradually but i have no idea if that would affect the outcome at all, because the problem with load could be that the servers can't handle that many users in the first place, even if you move them over one user per millennium
if you don't test, you're flying blind
so you SHOULD be worried
if you can't get a staging environment set up before you have to go into production, then you're just going to have to bite your nails through it and hope for the best
but before you do, you should probably write up an email (and save it in a safe place) to your boss telling them that you really think you should test the system in an integration (non-production) environment prior to rolling it out
350 users isn't peanuts... I mean on an Amazon EC2 scale it is, but it's not like, the kind of resources that one person would have in their house. this is a "real" IT environment. that deserves a testing environment imho
I'm sure you could write up a business case and at least get it to the principal decision-maker (CTO or CIO) to at least evaluate it
@allquixotic My environment is much more loose than that (which is quite nice really.) In fact, the whole thing could come crashing down and it wouldn't be the apocalypse. The old servers are still running and we'd just say "Hey click the old icons."
so you either give us a known quantity of money to buy and implement an integration environment, or you allow for an unknown quantity of money to flow out of our coffers while we are unable to conduct business until we restore services in the event of a migration problem
oh ok; so you're going to stand up the new environment in parallel with the old?
that basically is an integration environment then. XD just go for it and see what happens, and since you have the old environment as a fallback, you don't have to worry about lost business if it fails
we use integration environment for testing web applications on the public internet, because the "switch flipping" is basically changing the deployed application on the production URL (the only URL accessible to the public internet)
we have a separate environment that's an exact duplicate of the hardware and software of the new production environment-to-be, and the only difference is it's mapped to a different domain name for the purposes of testing
it's within our LAN and we deploy it exactly as we'd deploy to production if it were for real, so we can shake out any bugs that arise
but since for public-facing internet websites you don't have a notion of a fallback to an old version (in most cases), there's literally a binary switch you flip (whether it happens instantaneously or after a period of downtime is irrelevant) that goes from old version to new version, and there's no going back
so you ahve to be SURE that the new version will work. hence, integration testing
if your end-users are internal to the company, it's probably safe to let them access the integration environment and kick the tires (on a separate URL or domain name) to allow them to report any difficulties prior to launching it to production... although you'll probably want to have a fake database so that whatever changes are made within the integration environment don't affect real business data
but if it interfaces with the general public then you don't want them seeing your testing environment; you'll have to test it internally and flip it over during a downtime period after you're sure it works
you're kind of fortunate that your end users appear to be within the company so you can just stand up the new environment on a separate set of servers and let people migrate over "organically" as soon as they feel confident that the new environment is feature-complete, bug-free and usable for their work
general public end-users are much harder to satisfy and they often find bugs that you'd never in a million years anticipate during internal testing
anyway, I shouldn't trouble your thoughts with the problems that general public service providers like my enterprise have to face; I think based on what information you've provided me, you shouldn't worry in the slightest about your internal deployment as long as you keep the old servers around for a while.
It's a very mixed environment. Some locations we essentially "own", but under a separate company name. Other locations come to us for services. The latter I simply gave the option to migrate or not until we shut down the old ones. Luckily they're smart enough to realize that upgrading from a 2003 to 2008 environment is advantageous, but I kind of have to baby them with support calls since they can pretty much ditch us whenever they want. Today, however, Memphis has no choice. :)
the only other thing to remember is that you absolutely must keep track of which people are still using the old environment, and if they keep on using it due to some problem, it's your responsibility to contact them, find out why they're using the old environment, and make sure you either resolve their problems or otherwise convince them to migrate
@allquixotic Yeah that's gonna be a pain. A huge pain. I'll have to watch the list for a good week or two before we flip the switch on the old ones... A lot of people are scared of new things...
the last thing you want to do is kill the old production while people are still using it, whether it's because they're lazy, stubborn, or have showstoppers with the new environment that they either haven't had resolved, or haven't even mentioned to you
I'd also send multiple emails to your user base warning them that you're pulling the plug on the old production, weeks prior to doing so
These steps are kind of low-level and can be automated using, for example, NetworkManager on current GNOME 3 distributions. However, here are lower level instructions that should work across a wide variety of Linux distributions and versions:
Step 1: Get the usb0 internet connection working fine...
i feel like i did something wrong in that answer because i didn't test the steps, but i know them almost by heart so
nice. i was frothing at the mouth over the Xoom about a year ago when it was new. then i wanted the Transformer Prime. I ended up getting the ThinkPad Tablet earlier and the Nexus 7 now
so I have a 15" full-fat laptop with amazing specs (ThinkPad T530, maxed out); a high-end 10.1" Tegra 2 Android 4.0 tablet with a stylus; and a cheapo 7" Tegra 3 Android 4.1 tablet... and a monster of a desktop with a hardware RAID controller :)
now they need to invent dynamic matter creation technology that lets you press a button and automatically scale up a phone to be a tablet, or a full fat laptop, or a desktop
with a slider that lets you choose exactly how big you want it
they're a graphics technology company, with a long history of security issues. on top of that, forums are extremely juicy targets for stealing passwords and emails... so yeah, I'm totally not surprised that they got hacked
not sure I feel comfortable using their graphics drivers after how long it took them to fix the bug that allowed arbitrary memory writes through some device in /dev that regular users could access
@allquixotic then you might as well just have a giant computer hidden in foldspace and carry a keyboard/mouse/projector combo around with you, or hide them there as well...
I am disappoint! The expected arrival date for my new computer case (for my pending computer transplant operation) has been bumped up from today to tomorrow.
Not that I'd have the time to do anything with it today anyway, but that's besides the point. :P
- Fixed RandR per-CRTC gamma persistence across modeswitches and VT-switches.
- Fixed a bug that caused the X server to sometimes hang in response to input events. (Closes: #684941)
- Fixed a reduction in rendering performance for core X11 rendering on certain GPUs that occurred in the 290.* series of releases.
- Fixed a bug that prevented PowerMizer from working correctly on some boards with GDDR5 memory, such as some GeForce GT 240 SKUs.
- Added support for the following GPUs: GeForce GTX 660, GeForce GTX 650.
That reminds me... my home PC has been shutting off somewhat randomly. It's always the worst when I'm running a VM though. Yesterday it wouldn't go five minutes without an instant shut off when I had Ubuntu in Vbox. Does a little when I'm gaming too, or streaming video. Or maybe it was the dryer running... ~_~ What would you guys look at first?
RandR /care, dunno if I experience that hang (don't think so, previous other hang was fixed recently), wonder if my GPU was part of that reduction, don't have such board, support /care, don't have the bug, interesting that they enable FXAA by default (or maybe it imposes that it enables UBB when you enable FXAA)...
I have an associative array created by a type of rowtype of a table column.
To give an example, this is how it is(the table names are different, but the structure is the same):
This is the DDL of the table
CREATE TABLE employees
(
id NUMBER,
name VARCHAR2(240),
salary N...
I always thought it was overheating too, but after looking closer my guess is no. Though I have yet to pull any good clear and hard evidence on it, so it's time for that... Also there's I swear only two frickin' breakers in my house so everything is running on one of two lines. The fridge, microwave, and computers can't be on at once. Could voltage drops in the actual line cause this?
@Sathya: I totally know nothing about PL/SQL, but from a programming perspective a table can't be NULL given that the table actually exists, you might want to instead access the count of it. If this makes no sense, ignore my guessing... :P
Other than that I a gree with @allquixotic, what's the use of procedural SQL queries?
well, no, the idea is that you can migrate to a front end client of your liking, and your app logic can remain the same without redoing the whole thing
I'm wondering if we could broadcast data from a wireless access point in one place using a system in another network.
Suppose I'm in my home and I have to broadcast some data over the access point in my university. Is it possible???
Man, what a buzzkill. One of the GMs of the MMORPG running on my dedicated server just quit today.
And this is like a week before he told me he would pick up the cost of the server because he was so into it... amazing how quickly people change their minds
Guess I'm back to being the anchor for this game as always
@slhck It involves too much assumptions; so, in other words: Need more details!
I would assume broadcasting through a VPN tunnel could be a way to do it.
And I also assume this is for a project and not for any other malicious use...
We had someone take down the university network for half a day when we were implementing some distributed protocol.
And on a LAN party we had someone inserting both ends of a LAN cable into the router, dunno how that happened; took connections down due to loop back...
is that LAN port to WAN port, or LAN to LAN? Maybe I'm missing some knowledge but I don't see why a router would use a LAN port that's connected to itself?
There are obviously some more steps involved, but just to give an idea; I think a main focus is on keeping the path between in and out as short as possible.
Think that it's more reasonable to have an external circuit with relays or so that checks if the signal on the two cables is exactly the same for a few seconds and then disables them or so; doesn't introduce any overhead, although it costs more...
And putting implementation aside, it's perhaps better that the network actually fails than that it would be silent about it, it's not intended behavior for the cables to be inserted that way.
True, but if you have a ton of ethernet ports on one of those industrial strength routers and you have a bunch of them stacked vertically and horizontally, I can see how someone might misread a number on a socket and wire a switch/hub/router to itself...
and if you have some large number of (working) NIC connections and you fat-finger one of them and it kills the entire router... well... not good
I agree the best approach would be to have physical dedicated hardware that doesn't increase latency or add processing overhead to handle situations like that. Particularly on industrial/enterprise hardware
conditional statements are the enemy of low latency for networking equipment
Steps to reproduce:
Go to the excerpt history of a tag.
Click on edit.
Change the excerpt of the tag.
Save your edit and you'll see that "page not found" page.
I am trying to setup a Linux Mint 13 system using mdadm, with 2 1TB hard drives for a software mirror. I have tried following instructions found for Linux Mint 12, but the instructions fail to mention how to fix the Cannot install GRUB to /dev/sda error, even when I select the /dev/md0 volume as ...
It seems strange to me that every 7th generation console has much less memory than contemporary PCs, but games on consoles still run pretty smooth today.
For example:
PS3 has 256 MB memory for system and 256 MB memory for video.
Xbox 360 has 512 MB of GDDR3 RAM clocked at 700 MHz.
And, this is r...
Meh, perhaps consoles pulls it out of the site's scope a bit, but if you look at it the other way he's asking why a PC needs more memory to do the same task. Thus the question not really being about the consoles.
I have an associative array created by a type of rowtype of a table column.
To give an example, this is how it is(the table names are different, but the structure is the same):
This is the DDL of the table
CREATE TABLE employees
(
id NUMBER,
name VARCHAR2(240),
salary N...
I'm wondering if we could broadcast data from a wireless access point in one place using a system in another network.
Suppose I'm in my home and I have to broadcast some data over the access point in my university. Is it possible???
Steps to reproduce:
Go to the excerpt history of a tag.
Click on edit.
Change the excerpt of the tag.
Save your edit and you'll see that "page not found" page.
