« first day (626 days earlier)      last day (4641 days later) » 

00:33
Looks like the world has itself another Stuxnet: news.com.au/technology/…
01:17
I just died a little inside. Our sales guy is now calling our solution "cloud based" (which I guess is technically true, I just despise that term)
01:32
@MarkHenderson Have you tried pointing out that it is technically "silicon-based"?
@MilesErickson Hah, doesn't sound as good
Plus when I think of silicone I think of the stuff they use for sealing bathrooms
@MarkHenderson There's something wrong with a guy who hears "silicone" and thinks of bathroom sealant before breast augmentation.
01:54
@ChrisS Yeah its called being married :P And really not into fake tits
@MarkHenderson So being married castrates you? Good thing I haven't been close to such a fate in my 30 years...
@WesleyDavid No, but I found myself watching substantially less porn
And just generally less interested in other women
02:09
@MarkHenderson I can see that. Although, the same thing happened to me when I discovered IT as a profession. Not sure what to make of that.
Congratulations, you're the first person to mention that in Comms Room chat: http://chat.stackexchange.com/search?q=%22breast+augmentation%22&room=127
The other room to mention it was The Bridge. Surprise.
@MikeyB Sweet! Where's my badge?
@MarkHenderson You've never felt fake have you.... That or you mean you're not into overstuffed appendages that make most watermelon jealous. If the latter is the case, I completely agree with you.
Ouch, Adrian got a beat down.
@ChrisS No I've never felt fake :p
But I have felt large and small and large ones aren't nearly as much fun as you might think by watching the movies
02:45
Heh. I've not ever seen a Linksys that I trusted. I've dealt with so many "home routers" that I KNOW what utter shit they are.
@Adrian Yeah, I avoid them if at all possible. And if not possible, I avoid the potential customer.
The newer stuff has been quite good. The older un-flashable stuff was irritating in its TCP setting.
My GF's Linksys DDWRT has been a total CF. In fact, I'm once again trying to get it working, which is why I'm online right now instead of on the couch sucking down Cabernet.
@Adrian: the wrt54G?
wrt54gl
I'm starting to wonder if the older models were better made
02:49
v1.1
/me has a GV4
that is to say the GL before they split the line
they were. There's less memory in them now.
I might replace it anyway (eventually), since i'm getting to the point where most of my systems do gig-e
@Adrian They're good for DD-WRT and that's about it
@MarkHenderson And I'm not sure about even that anymore. This one sucks.
02:53
When Cisco bought Linksys we used to get the occasional Linksys customer who had called us instead. We then had to explain that they had a different support number to call and they would crack the shits because "FUCK I just spent 45 minutes on hold there and I gave up and called you! Now I'm at the back of the queue again!"
If it gives me too much more trouble it's gonna end like a clay pigeon and get chucked and shot.
@Adrian I have a small pile of dead WRT54G's :p
Most of them to be honest probably just need new power adapters
But they have such tiny specs you can only get the basic DD-WRT on it, which doesn't have any of the good stuff like routing protocols or VPN support
@MarkHenderson Heh. Old ones were good. I've been told that the ones since about 2006 have less memory.
i feel that way about netgears
was it netgear?
@Adrian Yeah they do. I have the v3 which has 1/2 the memory of the v2 :( And I think from v4 and above you can't load DD-WRT on them without a lot of work
02:55
no, dlink
Well I used to have. Recently replaced it with a normal N-based belkin AP totally un-flashed
@MarkHenderson yeah, that and the DIY nature of DDWRT on those things. I really can't recommend that route to small branch offices like that.
@Adrian No, not for actual corporate use
I love the tiny Watchguards for branch offices. Don't have any here but another client has them
They all "phone home" and pull their configs from a central store
@MarkHenderson Unless there's full-time on-site IT staff, that shit just needs to run: serverfault.com/a/393420/73435
New office? No worries, drop it into the IPSec diagram and every device automatically initiates a connection to the new office
02:57
I got a little pissy, but I think I stayed in bounds.
@Adrian FVS318 doesn't do IPv6 does it?"
We had an older FVS and its IPv6 was... lacking
@MarkHenderson: neither do dd-wrts ;p
Heh. You're probably right in that respect. Of course, some of that IPv6 seems to be getting past the routers and straight to clients judging on by what I'm seeing.
it's really strange.
@Adrian: probably tunneling
I've got it tunneled over UDP in some of my systems since it seems to sort of work, no matter what
02:59
Time to go. The Lady says I need to start the grill. Laterz.
it might also be teredo (which needs to die. Painfully. In a Fire.)
@JourneymanGeek Thats sad. I have an 8-year-old Snapgear device that supports IPv6
@MarkHenderson: no one really seems very enthusastic about it
It even has a built in Hurricane Electric thingo. One checkbox to enable and 10 minutes later everyone has an IPv6 address through HE.
03:01
@JourneymanGeek Teredo is almost useless in Windows. It uses IPv6 native first, then IPv4, then Teredo. Chances are if your connection fails on the first two, Teredo isn't going to help
Unless it's an IPv6 only service that only serves AAAA records and you only have teredo
@MarkHenderson: teredo is useless in linux as well ;p
I have been sooo happy with my dlink-825 or somesuch. Totally stable running DD-WRT.
The Cisco 610N WAP piece of shit on the other hand, ready to see if it'll blend.
And the WRT54GL, well it is what it is.
i've messed around with it for ipv6 access and its performance has been horrible
@MikeyB: maybe they got better, i have at least two identical d-links i know of fail mysteriously and suddenly, one of which had pathetic signal strength
Wee. Chat on the phone works tho. snicker
damn, the debian text installer has much more options than the ubuntu one
including a minimal init
03:05
@JourneymanGeek I suspect that for that kind of consumer-level hardware, you're rolling the dice a little. But yeah...
Oops, it was a DIR-615. (Is it bad that I went to my old SF question mentioning it to check the model rather than the device itself? :)
What the fuck…?
@MikeyB: hm, one of the ones i have has had all its labels removed and gutted
03:07
Nearly the same drive, except 128GB model on top vs. 256GB model on the bottom. What kind of fucked-up performance graph is that? (the wacky orange line is maximum IO latency)
Anyone ever seen anything like that?
Ugh. This SSD stuff is downright strange.
Yeah, it can get really weird. Don't forget though, I'm pushing 6000 freakin' random write IOPS through these drives. Trying to break them.
We have a customer who has an application that is VERY sensitive to write latency - if a single IO take 3 seconds to complete it'll pretty much shit a brick.
Client emails me about their MySQL database that pegged a quad core Opteron with 8GB of RAM during a recent traffic surge. "I read once that there was a setting in one of the configuration files that allows the database to handle surges better? To be more flexible?"
Other than adding make_mysql_flexible = true in my.cnf, I'm not sure what he's recalling.
2
Sounds like the Oracle stuff I worked on at Big Blue. Every device in the network reported in to our app, and it had a 10 minute data SLA including the report runtime.
03:27
I'm having a bit of trouble connecting to IPv6 addresses:
ping -6 2001:4860:b002::68

Pinging 2001:4860:b002::68 with 32 bytes of data:
PING: transmit failed. General failure.
PING: transmit failed. General failure.
PING: transmit failed. General failure.
PING: transmit failed. General failure.
I've checked the device manager to ensure the Teredo Tunneling interface is enabled.
How would I go about debugging this?
@GeorgeEdison: it seems to have issues with a native ipv6 box as well
@GeorgeEdison Do you uuh, have an IPv6 address?
Nope.
I thought the Teredo Tunneling interface would route IPv6 packets through an IPv4 network.
...or am I confused?
vaguely. in theory
it never works worth a damn ;p
How would I go about debugging it though?
@GeorgeEdison Can you ping a host that's up? 2607:f8b0:4002:801::1014 for instance? The one you gave doesn't respond at all.
I still suspect not as you're getting a transmit failure, but couldn't hurt. Personally I wouldn't bother debugging Toredo.
Same results, I'm afraid.
I fucking hate it when I get emails like this:
I am in financial meeting with Adam and Marisa and Adam has advised this needs to be fixed by the end of this month or there will be trouble.
Seriously, you have an oustanding issue for 3 weeks but you don't log a support ticket for it and then crack the shits when it's not fixed?
Especially when at first glance it seems to be a PEBKAC problem
@MarkHenderson Backhand time!
@MarkHenderson Am I lucky that I could take something like that to the person's boss and have them chewed throughly before they get it fixed?
03:59
@GeorgeEdison: windows or linux box?
Windows box.
I have Ubuntu running in a VM... I wonder how complicated it would be to tunnel IPv6 through VirtualBox's virtual network interface and then through my actual network :P
I hate that SQLSERVERAGENT is in all capitals. It's like it's screaming "FUCK YOU I'M A SERVICE!" #sysadmin
@ChrisS This guy is the boss
And they're our 2nd largest client
@ewwhite Do you use any software to tally how much to bill clients for the use of any of your appliances?
04:10
@ChrisS He's pretty good 99% of the time, but sometimes he gets a bee in his bonnet. Usually if we havent visited in a while and he might not feel like he's getting value for money (they're a 90 minute drive away from us)
Usually a quick phone call to remind him that we're still working (we deal with them probably 30% of a working week, but not with him directly)
@MarkHenderson Man, sucks. So you just brownnose your way out of it then?
@WesleyDavid Nah you don't even have to do that. Just instead of hitting reply, pick up the phone and give hiim a piece of your mind
I started to get pissed off at him for giving us an unrealistic 3-day deadline for something that his staff didn't log and after about 10 seconds we were all cool
@MarkHenderson Sounds like a reasonable fellow.
@WesleyDavid He is, absolutally. And he spents a shitload of money with us
I don't know exactly how much, but I think it's about 300k/year
04:16
@MarkHenderson That's nice. Is it all hourly MSP style stuff or do you host is infra?
Why the heck can't browsers just release a list of the parts of SVG & SMIL that they DO or DO NOT support?!
@WesleyDavid They host their own infrastructure, but we do the majority of the maintenance. We also developed their core LOB application.
And their system grows at about 30Gb/month during a slow month, so its a pretty busy system
@MarkHenderson sigh Sounds dreamy. I've come to the conclusion that I'm really not cut out for small business consulting where you're under the gun to either wipe and reload or troubleshoot to a successful end all within half a day simply because they can't afford your time.
I'm looking for options to level up to a different market.
Working right now with a nice guy that has an online, eCommerce site that's not doing too bad. But still, it's just a one man shop. Troubleshooting is minimal and reactionary, thus the server is starting to look like the Beverly Hillbillies car.
Yeah well it's not all roses :P The demands are pretty high. If we fuck up they start losing money, big time
Eek. Yeah, that's not fun. Ultimately, I've got different plans, but getting there is slow.
04:22
@WesleyDavid For appliances, no... but Observium can do it for per-port metering...
@ewwhite Gotcha. I didn't know if you billed for like "per 100,000 messages a month" or "per X amount of transactions" or whatever.
@WesleyDavid $180/hr... and anywhere between $100/mo and $300/mo for data storage.
I'm looking into how to tally per client usage of services. Basically, some kind of gateway / invoice tracking system to consume SNMP or MySQL data and then auto-generate and send the invoice or coordinate with the gateway for autoamted payment.
what types of services?
I like the flat rate approach... keeps billing simple.
@ewwhite Like percentage of emails relayed (smarthost), storage for hosted backups (Unitrends or rsync/rsnapshot)
04:26
Flat pricing for me... so my biggest client uses ~800GB of storage on my systems... they pay $300/mo.
But then, I have around 20 customers doing that...
@ewwhite Not bad at all. Is that file storage or something more specific like email archiving, or backup storage?
for smarthost services, it's just a freebie..
It's tied to the specific ERP application they use... so it's app-level replication on 10-minute intervals... think log-shipping.
But the clients also have images... tiff files of scanned documents. And I'd like to save myself the worry and move that to S3.
@ewwhite Yeah, but this might be bulk. Like there's a guy that I know who needs help sending emails to 400,000 subscribers to some bizarre game that apparently went viral in Asia / western pacific rim. So For customers like that, I'd handle outbound mail with like Exim and some baysian, ClamAV, CommTouch filtering and other magic to keep them from being blacklisted when sending their newsletters to subscribers.
so I'm trying to understand s3fs and s3ql as possible options.
@ewwhite Ahh, I see. Okay, interesting. Yeah, I'd do that kind of stuff too if I could find interested clients. Lately backups have seemed to be a concern, so I was looking at a Unitrends appliance and then do a monthly fee for backing up full client systems. Graduated pricing simply based on amount of space used. Doesn't matter if it's filesystem level, bare metal, or with fancy apps like Exchange or MySQL.
04:36
@WesleyDavid backups are important.
4
@ewwhite Starred for truths.
Restores are importanter.
But I'll admin that I know very few organizations that get good backups.
meanwhile, I'm back on the 3-host vSphere cluster...
@ewwhite I know. =/ It's so, so hard to get and maintain an excellent backup. It's the unsexiest job in IT.
83 VMs, SATA san with 14 disks in RAID5
@ewwhite Okay, I was mistaken. That is the unsexiest thing in IT.
04:39
and I don't think they're getting backups either.
Actually, I see Veeam running.
@ewwhite Thats the ultimate SAN gangbang
"83 hulking VMs. 14 disks. 7200 RPM. One party."
@MarkHenderson You forgot the cup.
It took 6 hours to copy 250GB between two VMs on the same host.
copy was limited to ~15mb/sec
@WesleyDavid No cups in this party. This is a normal gangbang that everyone can enjoy.
What do I tell the client?
"Just do this over... scrap it and start from the beginning"?
04:45
@ewwhite That about sums it up.
That's a tough spot to be in. You're the harbinger of doom and are essentially telling them that they either suffer in pain indefinitely or spend lots of money to do it they way it should have been done in the first place.
Well, 14 disks in a RAID 5... what's the technical argument against it?
I'd come armed with copious graphs and numbers.
1) How it's currently designed
2) How it should be designed
3) Consequences, pro and con, of each decision.
@ewwhite First is just speed. From my understanding of the communication, SAS at 7200RPM will edge out SATA at the same speed.
Well, they're in a bind. In the rush to virtualize, they put everything on a trio of old servers... pre-nehalem
Then, realize that 15K disks with proper cache will more than double performance.
@WesleyDavid Yes, queuing and error recovery.
04:49
Then, RAID five is a bit dangerous for such a large array.
I don't know how big their drives are, but I hope they're not bigger than 500GB
And I hope they have hot spares.
>_<
Are they at least "enterprise" drives? Like the WD RE4?
I've already had disk failures here... everything goes to crap...
04:50
With a higher URE number?
Whatever Dell specs
Okay, so that sounds promising at least.
Not like someone raided Best Buy for all the WD My Books on the shelves.
Rebuilding 1TB on such a contentious array will take days.
And will kill whatever is trying to run simultaneously.
Is it that the write speed is essentially that of one disk?
@ewwhite The write speed is essentially much less than one disk as the controller will be contending for the disk. See how the controller schedules its rebuild operations. It might be tuned for either aggressive or submissive.
IF it's aggressive, of course that'll take precedence. If submissive, the build will take a week or more.
@WesleyDavid well, general write speed. Assuming healthy array.
04:53
And then, with 1TB drives, your URE risk is higher. SO you could hit a bit that's just toast. Better hope it's not an important bit.
For random writes.
so it's a mess. I offered them a Nexenta array as proof of concept.
If you've got RAID6 and a good controller that scrubs platters, UREs can be caught and corrected before a rebuild finds them.
@ewwhite Did they slap that down?
@WesleyDavid Yes.
@ewwhite Yeah, sounds bad. So present your technical findings, submit your invoice, and run.
Oh, can't run... I support the ERP application.
04:56
IF they're balking at a NExenta box for what is a rather obvious problem with a rather obvious solution, then they're not serious about doing business and you'll do well to get your invoice in and...
...ohhh, shoot.
But here's the challenge... this is a place with internal IT people.
and the virtualization project was a big deal for them... so they wanted to run with it.
So I'd be kind of an ass to tell them that their ideas were off...
@ewwhite Wow, that's tough.
Wait, they're internal IT and they can't figure out what the problem is??
7200RPM array, with 80+ VMs, contended as heck
"OHAI!! Stuffs be liek, all slowz and stuffz!! IS IT CAN BE TROUBLESHOOTING TIEMZ NAO?!?"
I think they thought, "SAN + VMWare and a lot of RAM"
@ewwhite Ergh.
Maybe I'm a pompous ass, but whenever I do something, I read the HECK out of it to at least get a basic idea. And I think that just a few days of reading the heck out of this topic, built on top of what I consider to be prerequisite knowledge to calling one's self an IT person, should have told them that this was a bad idea doomed to failure.
You don't have to be a storage guy to know this stuff. Just basic hardware knowledge coupled with some targetted reading and chatting with people for a few days.
Well, it's one thing to be a consultant and see a lot of environments...
versus ONE environment.
05:04
Yeah, but... I dunno. I haven't seen a ton. I've been stuck contracting mostly for one place. I might as well have been their internal IT guy. But... maybe I'm a pompous ass.
but more than anything, I'm sad that they're using such old servers.
needed all sorts of VMWare EVC settings enabled.
@ewwhite I've heard some bitching about s3fs, but nothing concrete. I think the main problem is that it makes you think that S3 behaves like a filesystem (which it doesn't), and so stupid stuff ends up happening. Better to use something like s3cmd where it's obvious what network operations are happening.
@mgorven There's an s3ql that sounds interesting... I'm looking for an interface to use S3 as a WORM store for tiff image files.
@ewwhite Haven't used it, but it has some potentially useful features like compression, encryption and deduplication.
05:35
juju is very cool
05:54
@Tacticus Requires Ubuntu, though.
06:30
G'day
07:05
@ewwhite that's a good thing :P
cause frak centos
07:37
Morning all
@ewwhite I've used both s3ql and s3fs, It's not great.
One of them doesn't support UNIX flags/permissions (everything is 777) and the other is a expensive. If you change 1 bit in a block (which are 10 mb by default I think), it has to copy the role block over to S3 which is costly (transfer rate) and heavy on CPU (encryption and whatnot)
I had to design a backup solution 'in the cloud' as part of my thesis.
Dan
Dan
08:08
Morning
morning, 43 today, what an old bastard...
grats mate
I mean, grats - old fart
@Chopper3 many happy returns
Dan
Dan
@Chopper3 Hey, I'm 26 today. I wasn't going to mention it, but seeing as you have!
Congratulations to the both of you :)
08:24
Morning all. Happy birthday paradox day, you two.
@SmallClanger Ah, the pigeonhole principle :)
08:45
@Dan grats!
Dan
Dan
It's my own birthday and I still just looked at the system tray to get the date. Christ.
2
Thanks guys
@Dan I can remember exactly what I was doing on the day you were born, I spent most of the day in a pub called The Green Bank and was sick in a bin outside - nice
Dan
Dan
I had no idea my birth was so important to you ;)

Thanks guys :)
This guy confuses me... serverfault.com/q/393496/1435
@Chopper3 I bet he copied it onto something newer (like ext4), and that it created btimes
08:54
exactly...
3
Q: How to deal with mass out of office / company closure

user121246We have a 4 day weekend coming up soon and it has been suggested that all turn out of office assistant on to ensure all customers are aware we are closed. Due to end user reliability and stupidity i have been requested to come up with a way to set all of them. Is there a way to make a transport...

Due to end user reliability and stupidity i have been requested to come up with a way to set all of them. snerk but true.
Dan
Dan
@tombull89 Ah, the story of IT
Not a brilliant composed question but if someone answered this well it could be pretty canonical - I imagine there's a lot of that kind of request out there, could do a linux variant too
I guess he's UK-based, closed for 4 days because of the weekend and bank holidays.
08:57
hi, BTW, and happy birthday guys.
fanks
well, I thought clone was, y'know, an identical copy?
something might have gotten changed/corrupted
Dan
Dan
Surely easier just to re-clone
09:15
did anybody spot this on the BBC last week?
what should i be seeing?
@growse the "UNSC" logo is from the United Nations Space Command, in the Halo series. They meant to use the UN logo for the UN security council.
ok, that's very very funny
Dan
Dan
@tombull89 Ahh, Google Images - fucking over producers everywhere :D
 
1 hour later…
Dan
Dan
10:30
Wow, the Olympic web team have come up with the most pointless system in the world
You find your event, select the tickets you want. It then spends 15 minutes (Literally, no lie) 'reserving' said tickets before you can buy them. Except, well, it doesn't. So twice now I've waited, inserted my details and blam "Sorry, your tickets are not available"
My favourite bit is that, apparently, it's not even broken. The webpage just says "Sorry, yeah, those tickets are no longer available."
I've said it before, and I'll doubtless say it again, but fuck the olympics.
I've got no problem with the actual sporting events themselves, but fuck the whole "olympic" package.
I bet the bits I want to see (equestrian) will be mostly not shown or will be on at really weird times
Dan
Dan
@DJPon3 Despite having tickets to see something, I have to agree. The whole thing is far too overtly corporate and has lost the whole ethos of what it's all about really
Yeah that's the bit I'm talking about, the whole "corporate event" thing that we're having to spend a fortune on
want to watch the horse jumping? Watch it with my blessings (as if you need them, I know!)
got a ticket for something you want to see? nice one!
I hope to see some of the Judo on TV myself.
But all the BS surrounding those events? All the concerts, the opening ceremony, the flame (that's just some lame idea the nazis introduced anyway, IIRC)? Not interested, sorry.
10:47
claim to fame: a friend's fiance's brother is competing in the kayacking events.
cool
I was thinking about going to london for a week during the summer but the prices and the olympics have put me off a bit.
@DJPon3 as we have good prospects for showjumping and eventing we'll probably get them at reasonable times, tbh it's the dressage I'd like to see some of
Dan
Dan
@tombull89 Yeah, I'd avoid London at that time unless you're actively going for the Olympics. Transport will be hell - I've booked a hotel just inside the M25, based solely on the premise that I can get to Stanmore station and then not have to change tubes until I'm at the park
ah.. fair enough
10:50
@tombull89 a friends, cousin's horse is competing for New Zealand (eventing) in the olympics. It should have been competing for Ireland but the friends cousin was killed riding it at a cometition in Ireland
yeah my brother's a road warrior for a "green" heating company and they are dreading the transport disruption for their london sites
@Iain That's unfortunate - no sport should cost a humans or animals life.
yeah - eventing is dangerous
actually - standing next to a horse is dangerous and it gets more so from there on
Dan
Dan
@Iain Yeah, I like horses as animals but I prefer to keep my distance.
I've cracked an elbow and ahd a few bruises from falling off but my worst was broken bone in my foot from standing next to one and being stepped on
10:55
I'm going to go with 'ouch' there
Also, having just read the star wall, happy system tray day Dan
Dan
Dan
@DJPon3 Thanks dude, it's Choppers, too
Happy birthday to @Chopper3 too
@Dan Happy Birthday
Dan
Dan
@Iain Cheers :)
Yay, got tickets to the Orbit
11:18
@Dan what event is that ?
is this sophisticated spam, cluelessness or something else serverfault.com/questions/393560/… ?
Dan
Dan
@Iain It's just to visit that big tower thing they've built in the park
-1
Q: windows Server is hacked by SOG how any one please help me?

swapnil bhutkarmy server is Win2K8 Web SRV , IIS-7, it is a web hosting server & my sites was hacked by some pakistani terrorist groups it is hacked by SOG say's Soldiers Of God From Pakistani Group, please help me to fight the situation. how can i stop attacking from these groups.

Another 'O NOOOZ MY SERVER GOT H4XORED'
12:20
@Chopper3 Happy birthday.
5
Q: Convert a Smart Array RAID5 array to a RAID6 (ADG) variant

Pawel SawickiI have an HP MicroServer (N36L) with a SmartArray P410 controller, a 512MB BBWC unit and four (4) 2TB disks connected. Three of those disks are configured as a single RAID5 array and one is left alone. I'd like to migrate the RAID5 array to a RAID6 array. Is it possible to migrate it "in place...

if the guy read the instructions posted, he could have figured this out in the same amount of time it took him to take the screen shots.
Dan
Dan
I've been asked to confirm which version of ASP.net an IIS7 website is using - can anyone point me in the right direction?
@tombull89 That's awesome!
@Dan nmap -O? No, I'm not sure.
Dan
Dan
Ah, application pools, got it
@Dan Nessus should be able to find it too.
12:27
@ewwhite I've seriously thought about changing my resume, under skills, #1 Can actually follow directions #2 Strong Google-Fu, can find directions #3 Effective communications, can write directions.
There seems to be a shortage of sysadmins, particularly in small businesses, with those skills.
@ChrisS They're important skills... and I ask that of engineers I've interviewed and of potential team members that I want to join...
I mask it as "how do you solve engineering problems and what resources do you tend to use?"
The last 3 or 4 interviews I had, I asked the existing engineers that question. None use ServerFault... One guy even went as far as to say that the answers here were often wrong, so it was discredited in his eyes...
The answers here are often wrong?
Writing a good question is almost as important (communication skills, getting to the point, etc.)
@ewwhite I've found wrong answers before, so I guess it's possible someone ran across several of them. It'd be nicer if he corrected them, but no loss to him either way.
Wrong as in wrong, or wrong as in not what he wanted to do so he disagreed "wrong"?
12:32
@BartSilverstrim Yeah, another Linux engineer I interviewed with landed on the site a few times from google and was not impressed.
But didn't say what made them wrong.
@BartSilverstrim Basically... I suggested that he correct what he saw... and then went on to say that there's community pressure to correct bad answers here.
And he said?
I wonder what happens to the votes anonymous people put in on the site... I know it asks them if the Question was helpful: "yes" or "no"... Perhaps that data could help us clean up some of those answers.
@BartSilverstrim "Um, I have some good engineers in Australia to ask questions of... so I don't need to go to the internet that often to find answers."
12:36
ah. A cop-out.
@ewwhite I know that's why I and other people from my company don't ask Q on SF hardly ever, if we break something we just call MS and get them to fix it.
So it's a cultural thing.
@ChrisS I do it when I see something good on StackOverflow or Unix.SE
Well, I ask a lot of questions... but sometimes it takes someone else's experience to jog memory or to lead you in the right direction.
I spent 4 hours last night trying to get Linux LXC containers to work before giving up.
It may become a question here or on Unix.SE today.
How do you tell your client politely he's an idiot?
@BartDeVos I'm in that situation with the VMWare disaster I mentioned earlier...
You avoid saying "I told you so" and efficiently clean up the mess.
12:44
I can't seem to make him understand that spaces in a DNS-record (A/CNAME) are a BAD idea that I can't even implement...
my application.corp.tld will NOT work
But, he turned to google and said it was possible while he threw something on the table about DNS name spaces...
@BartDeVos That's one that's easy to prove, though.
ugh
yeah, showed him. "Maybe, you should be using another product then...."
Permalink to your vmware problems?
8 hours ago, by ewwhite
83 VMs, SATA san with 14 disks in RAID5
The client is running a Dell SAN with 14 7200 RPM SATA disks... but also has three ESXi hosts and 83 virtual machines running Linux apps, web servers, databases, Exchange and VDI...
12:50
And no backups?
I saw Veeam running, so there are some backups.
Nexenta would be a good call imo
@ewwhite Performance must be rockin'!
@BartDeVos I offered... but it's a case of the client not wanting to really admit the mistake.
@BartDeVos DNS supports it on a technical level, but no application supports spaces in domain names... So unless he can modify the app, which will be difficult to implement and support, he's still up a crick.
12:53
We run lots of vm's on S-ata.. works fine :)
@BartSilverstrim I can only image what a 1KB random write test would look like on that array...
@pauska SATA isn't the problem. It's the 14-drive RAID5.
@ChrisS Well, I'd like to show them a graph of how bad things are... what's the best vSphere graph to show? Queues? Latency?
@ewwhite I've never touched vSphere =]
@ewwhite I'll have a look
Throughput and Queue depth should show off "performance" though
12:56
Well...are they hitting the VM's enough that it shows an appreciable lag?
@BartSilverstrim Yes. I needed to copy ~200GB from a VM on the iSCSI RAID5 SAN to a new VM I created on local disks. The transfer was capped at ~15 megabytes/second because the read speed from the array is so bad.
THEY NEED MORE MEMORY
Dan
Dan
http://www.bbc.co.uk/news/uk-england-derbyshire-18250397

Jesus, someone must've been confident to go ahead and clear off that arrest
it should have been 80-90 megabytes/second (over the virtual 10GbE NICs)
@BartSilverstrim Lots of RAM, though.
Ugh, looks like our stats b0rked out (again)
once you have multiple data centers in you vsphere the thing goes crazy

« first day (626 days earlier)      last day (4641 days later) »