what did you learn today? (part 2)

[Rick25]

VAAI doesn't require a plugin AFAIK, unlike the multipathing plugins for storage isn't it automatically activated if the underlying storage supports it?

 

[Rick25]

yeah I'm really not feeling the webclient. I would really like them to save your location within the client into some sort of xml or whatever so that it would reopen to the same spot the next time. I hate having to dig down three levels each and every time.

 

[Rick25]

euri[/url]":1mh4avep]I like the Web Client and SSO just fine. Just not the versions y'all are using.

Someone's playing with the beta! Really hope they're announcing 6 next week.

 

[Rick25]

Ballmer's on his way out. Be very interesting to see what happens next. Though based on the stock jump (9%) the Street likes the news

http://finance.yahoo.com/news/microsoft ... 09911.html

 

[Rick25]

That a hardware vendor who's hardware decided to let go the magic smoke expects us to cover the duties at the border for the RMA unit despite us having bought the add-on warranty sku. Getting rid of that vendor's line just moved up my list of priorities before the next renewal.

 

[Rick25]

euri[/url]":2b1lb5hz]So, today I passed the VCAP5-DCD and found out I passed the VCAP5-DCA on the 12th.

Congrats!!

 

[Rick25]

That the issue with cleaning up the source mailbox during a mailbox move in Exchange 2010 (post SP2 RU3) got fixed with SP3 RU1!

 

[Rick25]

That Microsoft has pulled the Masters Program from their certification track
http://www.zdnet.com/microsoft-abruptly ... 000020093/

 

[Rick25]

TIL we did the ESX storage redundancy right. One of the 10G switches decided to boot itself. Aside from alerts from Ops Manager about the lost redundancy and huge spike in Disk I/O registering everything kept humming along. 4mins later all back up and running. Opened a ticket to figure out why. I'm assuming cause we're in the process of replacing it with a 48port switch and it someone knows about that.

 

[Rick25]

That P2Ving a couple of Cert servers isn't as smooth as expected. Also doesn't help that when you go back to the physical and power it on it doesn't get past the BIOS screen.

Thankfully a CRL publish and renewing 1 DCs UserAuthenication cert seems to have resolved the issues. It's 9:54a and I could use a serious drink.

 

[Rick25]

TIL that if you've got a DC that keeps disabling In and Outbound replication you've got a USN rollback on your hands.

Strange thing is it wasn't a restore or a VM snapshot. We had to move the VMDK files so the server was shut down, moved and restarted. May have
just been a timing issue.

Pulling it down and Metadata cleanup on tap for Monday.

 

[Rick25]

Server unreachable, system board had been replaced a couple of week back, 1 guess if iLO had been reconfigured.....

OS is frozen so power it down, then it's just doing a quick cycle thing. Remove power, wait and reconnect now won't even power on.....Open a ticket....

Metadata cleanup from the USN rollback on a DC from Friday

Motherboard from server is replaced and firmware updated....but the RAID controller doesn't want to bring up the logical volumes (they're unavailable). Couple of reboots and a complete power down later and it's all good.

Is it Friday yet?

 

[Rick25]

Danger Mouse[/url]":h35xl1ur]
In other news...

TIL that I have been promoted. A supervisor (in reality assistant manager or ops supervisor for all IT operations) is me.

Congrats, well deserved!

 

[Rick25]

Digitlman[/url]":2ffxzgwj]Some part of Google is broken this morning. It took forever logging into CDW, and the web page was waiting for google.com for something.

Now, logging into Ars, the login again took forever waiting on google for something...and now the main stories are gone from the bottom of the page.

Anybody else seeing this?

Is google hosting one of the new health exchanges and it is overloaded?

edit: could not even paste smiley code. And now the stories popped in. WTF? Is it Monday again?

Gmail is taking a long time to respond or is timing out for the last hour or so.

 

[Rick25]

100kW APC PDU frames are very very heavy. Not having a proper transition from hallways to datacenter doesn't make it any easier. +1 for Facilities folks with great ideas.

 

[Rick25]

TIL

a) That I love the new Gen2 APC PDUs. Love it even more now that we've upgrade four of them to 30A circuits. Think that's going to be my standard going forward

b) the spreadsheet that I built to turn Amp loads into kW worked pretty well. We were within a kW of the expected loads

c) that our UPSs have a very nice linear drop. So once we shut off Utility power we got a run time number that turned out to be pretty accurate.

d) we've got good electricians that got the new breakers in well ahead of schedule (little bit of Scotty based estimating there I think)

[edit]Corrected kWH to kW[/edit]

 

[Rick25]

llib[/url]":3uv9ymo4]

Rick25[/url]":3uv9ymo4]TIL

a) That I love the new Gen2 APC PDUs. Love it even more now that we've upgrade four of them to 30A circuits. Think that's going to be my standard going forward

b) the spreadsheet that I built to turn Amp loads into kWH worked pretty well. We were within a kW of the expected loads

c) that our UPSs have a very nice linear drop. So once we shut off Utility power we got a run time number that turned out to be pretty accurate.

d) we've got good electricians that got the new breakers in well ahead of schedule (little bit of Scotty based estimating there I think)

Very good, Mr. Rick. Very good indeed!

Thanks llib!

LordFrith[/url]":3uv9ymo4]

Rick25[/url]":3uv9ymo4]TIL

b) the spreadsheet that I built to turn Amp loads into kWH worked pretty well. We were within a kW of the expected loads

Did you turn them into kW, or into kWh? And were you within a few kW, or a few kWh?

kW because that's what the APC engineer could plug in to spit out an estimated runtime.

 

[Rick25]

Barmaglot[/url]":2v8hmsf9]Sanity check please: if I need to provide redundant power to a triple PSU system (that needs two of three PSUs to run), the proper way to do it is A feed to PSU1, B feed to PSU2, and an automatic transfer switch plugged into both feeds and PSU3, correct?

Yes, that's exactly the solution I came up with to feed our Eql 65x0 arrays. The APC ATS is great. What I did was A into PSU1, B into PS3 and ATS into PS2. We also color code our power cables. Black, White and Blue for ATS power.

 

[Rick25]

Barmaglot[/url]":2lpvcxic]Good We just ordered a Tyan FT77A, and there should be more following it if it works out well

http://www.apc.com/products/family/?id=14 Bunch of different options. I'd love to see what engineering logic made them pick 3 instead of 2 larger or 4 smaller power supplies.

 

[Rick25]

That there's a bug in the NIC firmware in the Type 14 Eql controllers if the "Broadcom 84833 Firmware Version" is about level 1.39. During a firmware upgrade it can cause the Active controller to stop responding on it's eth port causing the array to drop offline until the secondary controller has completed it's reboot and taken over. The controllers never lost internal communications but we lost outside access. Waiting to hear what the fix is for it.

 

[Rick25]

antiwraith[/url]":20w5ief0]

Rick25[/url]":20w5ief0]That there's a bug in the NIC firmware in the Type 14 Eql controllers if the "Broadcom 84833 Firmware Version" is about level 1.39. During a firmware upgrade it can cause the Active controller to stop responding on it's eth port causing the array to drop offline until the secondary controller has completed it's reboot and taken over. The controllers never lost internal communications but we lost outside access. Waiting to hear what the fix is for it.

You have a link? Freaking Equal Logic......

No link. Sent you a PM

In other cheery news 6.0.7 is apparently out......

 

[Rick25]

TIL to check to see if the 600v 175A input feed and 208v 225A sub-feed conduits are being installed into the correct spots into the PDU frame. Thankfully we caught it before too much work was done.

 

[Rick25]

sryan2k1[/url]":7kknhjrw].... he said "Great so far, but I just got here so I've got plenty of time for it to go sideways"

++ that's a great quote

 

[Rick25]

That my APC SE's are just awesome. Power and Cooling guys.

 

[Rick25]

I've got the new beta for the FluidFS just need to dig up the time to get it installed (Eql hardware). Any idea if the access based enumeration got enabled for NTFS permissions rather then just Share permissions?

 

[Rick25]

afidel[/url]":2eekpsr1]
I'd say the max 1500 clients would keep most shops from getting anywhere near 1PB.

Where did you get the 1500 client limitation? Just reading through the notes for 3.0 and at least for the Eql NAS it's showing 10,000 for single appliance and 20,000 for a dual. Max limit on the filesystem size is 510TB

 

[Rick25]

Hmm interesting. Wondering if that'll change when Fluid3.0 releases

 

[Rick25]

Generally they're all done at testing centers.

If you're VMware heavy thing I'd definitely start with the VCP5. Upside is that it's a single exam, the negative is that to be eligible to write the exam you need to have taken one of the qualifying official courses. Here's a good place to start looking http://mylearn.vmware.com/mgrReg/plan.c ... i=www_cert

 

[Rick25]

akro[/url]":11828h34]I thought we used to have an "OOPS Moment Thread" sounds like a good one to revive....

Yep viewtopic.php?f=21&t=85191 Some great stories in there

 

[Rick25]

That I really like the APC Symmetra 100kW units. They've made a bunch of changes vs the older 80kW frame and overall they're pretty good decisions.

 

[Rick25]

That in IT it comes in handy if you can speak multiple languages a) electrical contractor b) facilities (and building automation c) UPS and Datacenter d) Fire Suppression Systems

TIL that if you and Facilities don't have the warm and fuzzies about the successful bidder (lowest bidder) for the Fire Suppression semi-annual check contract that you should a) acknowledge the disturbance in the force and b) have Facilities escort them out

Thankfully they actually didn't cause the Agent to discharge but they won't be back if I get a vote.

 

[Rick25]

That apparently there are people who believe that Apple built in a Bitcoin mining application into OS X and that the command line way to turn it on is "sudo rm -rf/*"......

 

[Rick25]

YIL that's there a bug with Server 2012 and Hyper-V that can cause the Service account to loose "Log on as a Service" rights...predictably this causes your Hyper-V cluster to bite the dust.

http://support.microsoft.com/kb/2779204

MS still looking into the cause of it.

 

[Rick25]

blokey[/url]":190a7gqd]

You can't wild card different top level domains. He was talking foo.com bar.com and baz.com having a single wildcard cert.

Really? Says who, because we have several certificates issued from Verisign that do just that. We have <ourdomain>.com <ourdomain>.co.uk <shortdomain>.co.uk and others in 1 SAN certificate.
And Verisign even advertise that you can do just that on their website:
http://www.verisign.com/ssl/buy-ssl-cer ... tml/?print

For example, you can secure all these domains with a single SAN Certificate:
http://www.bizybank.com
http://www.exchange.bizybank.com
http://www.bizybank.net
http://www.new-bizybank.net

You're talking about a UCC certificate. They came into being during the Exchange 2007 days. It allowed you to put a bunch of domain names on a single certificate. Originally they were a great deal, 10 SAN names and unlimited reissues.
Then the Registrars got greedy, it dropped to 3 names and extra $ for every one you added.

Wildcard certs secure a full domain *.domain.com but you can't combine them with other domains e.g. *.domain.org etc. Don't believe that it'll work with *.child.domain.com either.

 

[Rick25]

From the Verisign (symantec...yuck) site

Secure Site Wildcard SSL Certificates
Protect the transfer of sensitive data on multiple subdomains under one domain on your server using a single easy-to-manage certificate.

Bolding is mine

Just an FYI, I picked 10 servers for 2 years and they gave a rough price of $32k....holy (#&(. I think I paid $700 for 5years from GoDaddy and unlimited servers.

The other thing is that the wildcard shouldn't have any SAN names attached, just the single *.domain.whatever

 

[Rick25]

TIL that upgrading a wireless card on a laptop isn't the cookie cutter operation that it should be.

Removed a centrino N-1000 card and replaced it with the Centrino 6235. Ping time to the router jumps from 1ms to several hundred, download speed on 5Ghz is at 2-3MB/sec, 2.4 drops to 700KB from 3MB.

Testing order in this case was 5Ghz and then 2.4.

Rebooted laptop, connected to 2.4 SSID, transfer speeds at 10MB/sec with ping at 1ms. Thinking the reboot fixed this, moved
over to the 5Ghz SSID. Back to a 2MB/sec download, flipped over to 2.4 (no reboot) and I'm back to high ping times and very slow speeds.

Reboot, stick to 2.4 and I've got 1ms and 10MB speeds again.

No idea what would causes that sort of behavior, but it looks like I'm stuck on 2.4 at least for now. I've got the most current drivers too.

 

[Rick25]

Nope, the 5Ghz shows up just fine. Can see it and connect, it's just the performance that's in the toilet.

 

[Rick25]

Apparently there's a really long thread on the Intel Support forums about issues with this chipset. I'm running the drivers that are supposed to fix most of the problems. Only AP handy to test with unfortunately.

 

[Rick25]

That our ISP forgot 'wr mem' when they finished their switch config. Dropped the rack to plug the single PDU into a transfer switch and dual feed it. Switch came up ok, not so much for passing traffic though. Couple of workarounds for DNS and routing traffic out our other ISP until we could get the vendor to redo the config.

 

[Rick25]

dredphul[/url]":2c980bna]C level execs are talking about requiring 3000 days without significant (aka hours) downtime in production.

Given that's around 8.5 years , how many 9's of uptime is that?

May need to provide ammo explaining that uptime requirement can't be guaranteed without significant expense and is an unrealistic requirement.

Counting or not counting planned outages (patch/upgrade windows)?