ok now Dell tech support are having trouble sending emails out from a part of their website.
I must remind our dell account manager of that the next time he rocks up and says "I know you've got lots of email server knowledge on site already, but you should still buy our email server consultancy package...."
the admins should do what management says (as long as management understands whats being requested) within reason. management says its ok to have access - give it to them
1. I'd remind the Devs they can (perhaps "should") be fired for such crap 2. Tell them to document when they put requests into Engineering and when Engineering responds. The first things that's delivered late pull out the documentation of what's going on.... This is basic CYA in business land, something they should know already.
I don't think management knows... management is deferring to the engineering team to set policy.
This is a company that has gone through SIX people in four years to replace me. Each replacement brought stranger and weirder policies to the environment.
I can't lie, occasionally I'll 'work around' things if it's in my remit and it's easier than asking, but if something is being actively denied then yeah, I'll simply sit back, send some e-mails and explain that someone is now paying for me to read the Internet
as administrators, that seems key... make the devs happy... especially because that's usually the main part of the business.
@basil I have a discussion planned with Racktop Systems today. They are a Nexenta partner that OEM's their own hardware. They want moar marketshare and I want to grill them on how they can compete with proper enterprise storage.
I find Dan's method is the best way of getting things to change. It's maybe not the best way to "get things done" today, but it's the best way to make sure you'll be able to get things done tomorrow and the day after.
The question really should be put to management - why they're letting the engineers set policy, when the dev's are the ones doing the working and bringing in the revenue. Engineering should be supporting the devs
@basil I want to understand some of the characteristics and indispensable features that are common in the storage systems you manage.
@HaydnWVN I think it's fear. The management is dev-centric and never took time to understand systems (which is why I had a hard time there). As a result, they blindly trust the engineering leader.
The devs told me that the engineering team has a "no download" policy in their new system implementations. They went from CentOS to paid RHEL and no longer want anything downloaded outside of the RHEL repositories. No rpmforge, no epel... Yet, they're not creating custom RPMs for the things that are needed.
Tomcat is one of the applications... and there's a .so file that's no longer distributed with the RPM for mod_jk. Engineering says that's no good... "Don't use mod_jk anymore"
@ewwhite Management - whatever their situation always favour 'if it ain't broke don't fix it', particularly if the engineering manager has been there longer (= more trusted), but it's when things are causing additional work for the devs that it needs to be highlighted
@ewwhite The deficiency is that you have software from one place, LSI raid hardware from another place, ZFS logic from a third place, firmware on the HBAs, and who-knows-what handling failover and failback. If it stops working, how long will it take to fix? Who will I call?
@ewwhite Netapp is considered enterprise the same way 3par, compellant, or EMC's entry level box is. It's got one group of people in charge of making it all work together, but it doesn't do mainframe or support five nines.
@Basil That's my first question to this vendor... because I KNOW this thing can outperform the lowend EMC, Netapp and Dell solutions I see out there... but support means that the vendor would have to have very solid ZFS expertise on-staff and hardware techs to dispatch
@ewwhite Also, you'd be right for performance per dollar probably, but any of those other solutions are efficient and scalable in terms of performance.
You might find better performance per unit of hardware on something like compellant
After running following command:-
iptables -t nat -I POSTROUTING -m state --state NEW -p tcp --dport 25 -o eth0 -m statistic --mode nth --every 5 -j SNAT --to-source 173.224.222.45
I get following error:-
iptables: No chain/target/match by that name.
I am running this command to chang...
@SmallClanger He wants iptables to automatically pick a different external IP for each outgoing SMTP connection.... Smells heavily of spamming, though he does at least recognize that and mentioned it in the Q.
@OliverSalzburg Now you get to do my favourite thing - the non-accusatory, friendly, diplomatic e-mail which is secretly very accusatory and full of "told ya so, dickhead"
fun fact: in french, a crane is called a "grue". Every time I see a crane downtown for washing windows or whatever, it's got the word "grue" written all over it.
fun fact: in french, a crane is called a "grue". Every time I see a crane downtown for washing windows or whatever, it's got the word "grue" written all over it.
@Basil Yes I know the french are strange and feel that they can avoid being eaten by calling random things "grue", but why are we talking about grues? :-)
(fun fact: Grues won't eat the French as a staple diet because of all the butter -- too much cholesterol is not good for a grue.)
Heh (from: adobe.com/uk/products/creativesuite/faq.html#subscriptions) "Your subscription includes access to the current and future versions of the Creative Suite application (as long as your subscription remains active) as well as 10B of storage space and the ability to sync your creative files from your desktop to the cloud.*"
Probably not... I'm close enough to Chicago that I'd probably be feeling that kinda fallout too. 'Course, for how boring a Monday it's been, I might not mind some radioactivity and melting flesh to spice the day up a bit.
@ewwhite Ugh. My inner smartass loves to answer those... unfortunately.
I see the broken camera faerie visited one of our remote offices this weekend. "I took the camera out this morning to take a photograph and found the screen to be broken and the camera inoperable."
When the VMWare box is busier and coming out of an idle state, the 'Wait' performance metric drops down from a steady 140000ms "wait" time to nearly zero, and then goes back up.
the root of this; a custom piece of software we have on these thin clients likes to crash in the middle of the night - so their solution is to reboot all of them, every night, instead of figuring out why the software is prone to crashing
YAY! Our Printer Wizard is here! Couple times/year we stack all our broken HPs LaserJets in a corner and have our local HP certified technician come fix them all.
I get a little tired of people wanting stuff now and suddenly needing it so badly that it overrides all architecture tasks, but then whining when the stuff that gets delivered doesn't integrate well with the architecture. Duh?
@adrian Nah, that's because computer stuff is all just easy magic, and if you can't do both at once, it's because your magiks are weak, and you should be replaced by a more powerful computer wizard. Everyone knows that.
@HopelessN00b Heh, no. These are mostly social workers. It's all scary black magic to them. They just don't bother to make up their mind about anything until 5 minutes before they absolutely need it.
@HopelessN00b I was blessed in many respects. All my other work has been in software development companies. No too many people there for whom computers are black boxes.
My kids are getting big enough that I have to watch that now. Littlest is 50lbs and 4'4" and it's getting dicey regarding having her sit on my shoulders.
I told him that I valued expertise and that was why I came to him directly instead of getting the run-around from a generalist. He said he didn't change his own oil either :D hahaha
@adrian AD's just a hideous master-master networked database conforming to the whatever the hell MS felt like at the time standard, used for storing virtually everything about a Windows environment, including AAA functions.
Aye. I'd like to start leveraging global GPO so my padawan isn't having to set GPOs up for each one of our slowly-proliferating Windows terminal servers
If there's enough files to be a performance hit if grouped together and they're being called up by another app that dynamically generates the directory to grab, I see no problem.
Yeah, it's a situation where I want to start pruning years of images off of the system... and thinking that 6000 images per directory was a little stupid
@voretaq7 yeah, ATT started whining when the app would process backlogs and have to parse directories with 1M+ files in them. Took forever and they've have to move them to another box and run the grouping scripts there.
@ewwhite Can you discuss what the heck that thing does?
@voretaq7 Yeah, I have no idea what FS it was running. I was still pretty wet behind the ears back then and still figuring out non-Linux Unix-like OSs.
I especially like how the story of the filesystem is basically "So we doubled the block size again, and got about a 100% performance improvement (again)"
Check my previous question.
I am now looking at specifics relating to enterprise virtualization. I have done some research into Citrix Xen Server and VmWare vSphere. I need to know whether it is possible to run Windows 2008 guests on linux hosts. I would like to set it up like this for cost reas...
@ewwhite So you're going to be using the Cisco VPN client? In that case you have to choose the authentication mechanism, AD doesn't support them all. If you're FIPS-compliant it's a whole different world of certificate pain. Are you going to use per-client certificates?
That's actually the first time I've seen this particular BSDCan talk come up in my searches so I know what I'm listening to tonight after work. Everything I know about UFS journaling is off the mailing lists and out of the code :P
After running following command:-
iptables -t nat -I POSTROUTING -m state --state NEW -p tcp --dport 25 -o eth0 -m statistic --mode nth --every 5 -j SNAT --to-source 173.224.222.45
I get following error:-
iptables: No chain/target/match by that name.
I am running this command to chang...
