what did you learn today? (part 2)

[Danger Mouse]

...that I really love:

http://policelli.com/blog/archive/2011/ ... -topology/

The Microsoft Active Directory Topology Diagrammer

http://www.microsoft.com/en-us/download ... x?id=13380

 

[Technarch]

WingMan[/url]":w707vh35]

Altiris and work properly are oxymorons. It's gotten worse since Symantec took it over.

++ Pre-Symantec, we tried setting up Altiris in the lab as a trial. It just wouldn't work. We called the Altiris rep and said, "It's not working." Rep said, "Okay." No attempt to get us support or anything, it was like he knew it was doomed to fail.

Supposedly there are working Altiris installs in the city, but I haven't actually laid eyes on them.

 

[gblansandrock]

Danger Mouse[/url]":30qf78py]...that I really love:

http://policelli.com/blog/archive/2011/ ... -topology/

The Microsoft Active Directory Topology Diagrammer

http://www.microsoft.com/en-us/download ... x?id=13380

That sounds awesome DM. Thanks.

 

[hawkbox]

Danger Mouse[/url]":3km283d8]

euri[/url]":3km283d8]There's some support for RTL chipsets in later builds of ESXi. If I'm not mistaken, at least the 8111E works.

http://www.kendrickcoleman.com/index.ph ... ere-5.html

That would support your assertion. I would expect the 8111E would work, but earlier models not so much.

ESXi v5 and later should support it, but no guarantees.

http://communities.vmware.com/message/2067708

Realtek 8168
Realtek 8169
Realtek 8111E

No 8100A/B/C probably

--

Blacken,

if you need an Intel gbit NIC, pay for shipping and I'll gladly shuffle one off to you. I can probably find you a dual port NIC if so desired.

PCI-E V3? Cause if so I will pay for shipping a dual port nic. I have a quad port pci-e v1 that doesn't work in my T320.

 

[Blacken00100]

Danger Mouse[/url]":2nwojtrk]if you need an Intel gbit NIC, pay for shipping and I'll gladly shuffle one off to you. I can probably find you a dual port NIC if so desired.

Thanks, I appreciate it! Let me do a little more poking around to see what I want to do with the home lab stuff and I'll let you know if I need to go that route.

 

[Danger Mouse]

hawkbox[/url]":1x3vvtqg]

Danger Mouse[/url]":1x3vvtqg]

euri[/url]":1x3vvtqg]There's some support for RTL chipsets in later builds of ESXi. If I'm not mistaken, at least the 8111E works.

http://www.kendrickcoleman.com/index.ph ... ere-5.html

That would support your assertion. I would expect the 8111E would work, but earlier models not so much.

ESXi v5 and later should support it, but no guarantees.

http://communities.vmware.com/message/2067708

Realtek 8168
Realtek 8169
Realtek 8111E

No 8100A/B/C probably

--

Blacken,

if you need an Intel gbit NIC, pay for shipping and I'll gladly shuffle one off to you. I can probably find you a dual port NIC if so desired.

PCI-E V3? Cause if so I will pay for shipping a dual port nic. I have a quad port pci-e v1 that doesn't work in my T320.

PCI-X or PCI only. I can't give away actual working kit in production.

I can give away stuff that even *WE* wouldn't use in production.

 

[hawkbox]

Danger Mouse[/url]":2lm721d3]

hawkbox[/url]":2lm721d3]

Danger Mouse[/url]":2lm721d3]

euri[/url]":2lm721d3]There's some support for RTL chipsets in later builds of ESXi. If I'm not mistaken, at least the 8111E works.

http://www.kendrickcoleman.com/index.ph ... ere-5.html

That would support your assertion. I would expect the 8111E would work, but earlier models not so much.

ESXi v5 and later should support it, but no guarantees.

http://communities.vmware.com/message/2067708

Realtek 8168
Realtek 8169
Realtek 8111E

No 8100A/B/C probably

--

Blacken,

if you need an Intel gbit NIC, pay for shipping and I'll gladly shuffle one off to you. I can probably find you a dual port NIC if so desired.

PCI-E V3? Cause if so I will pay for shipping a dual port nic. I have a quad port pci-e v1 that doesn't work in my T320.

PCI-X or PCI only. I can't give away actual working kit in production.

I can give away stuff that even *WE* wouldn't use in production.

Lame. :devious:

 

[w00key]

Realtek was always kinda horrible. Unexplainable problem somtimes just vanishes when you switch to an Intel card, or when you upgrade, downgrade drivers and fiddle with options until it goes away. Nowadays I just insist on either installing an Intel® 82574L based card (€25) or having it onboard (Supermicro seems to like them).

 

[Matt Wallis]

w00key[/url]":11ny1t9h]Realtek was always kinda horrible. Unexplainable problem somtimes just vanishes when you switch to an Intel card, or when you upgrade, downgrade drivers and fiddle with options until it goes away. Nowadays I just insist on either installing an Intel® 82574L based card (€25) or having it onboard (Supermicro seems to like them).

Yeah, I keep a couple of Intel cards and even an old 3COM card floating around for fixing issues with computers running Realtek cards. The cards are cheap enough. I remember having a lot of trouble with early Broadcom chips as well, although they seem to have cleaned up their act.

 

[SandyTech]

PCI-X or PCI only. I can't give away actual working kit in production.

I can give away stuff that even *WE* wouldn't use in production.

Is it kinda scary that we're running stuff in our production environment that even DM's .org wouldn't? We've still got pair of ML350 G3s in production, and the RAID and NIC cards are in PCI-x.

edited to fix quote

 

[M. Jones]

sryan2k1[/url]":2uihpri4]I'm really not sure why anyone would suffer through the horrors that is Xen when ESXi is free.

Free in English can mean gratis or libre or both. ESXi is gratis, but it still requires a license key, even if that license key is gratis.

 

[sryan2k1]

And 99.9% of people won't care that you have to log in to the VMWare portal to get your free key. Actually, you get the key before it lets you download the ISO.

 

[Rick25]

Danger Mouse[/url]":agrhns94]...that I really love:


The Microsoft Active Directory Topology Diagrammer

http://www.microsoft.com/en-us/download ... x?id=13380

We've been using that for years, great program!

 

[Danger Mouse]

Matt Wallis[/url]":35j17d90]

w00key[/url]":35j17d90]Realtek was always kinda horrible. Unexplainable problem somtimes just vanishes when you switch to an Intel card, or when you upgrade, downgrade drivers and fiddle with options until it goes away. Nowadays I just insist on either installing an Intel® 82574L based card (€25) or having it onboard (Supermicro seems to like them).

Yeah, I keep a couple of Intel cards and even an old 3COM card floating around for fixing issues with computers running Realtek cards. The cards are cheap enough. I remember having a lot of trouble with early Broadcom chips as well, although they seem to have cleaned up their act.

Yeah, the drivers were horrible also. Heck, even Marvell has more or less come around.

Realtek seems to finally have cleaned up its act.

I'm just waiting for them to drop a bomb on other network vendors by dishing out a $20 10gb NIC

----

In other news, want to slow down your Exchange 2010 DAG and have random failures of replication and even storage presentation within the VM?

Easy, instead of doublechecking the NIC MAC addresses, cleverly barge on ahead assuming that all 4 of the VMs are properly configured instead of say...2 of them And those two misconfigured VMs (just wrong IPs) have similar but opposite failings.

I had a VERY hard time believing that MSCS was this crappy, so I knew it had to be a misconfiguration somewhere. Blam.

PEBCAK. I was in a rush to get it done (political pressure) and mistakes are the result.

Working good now.

In other words, when you're sure it's not PEBCAK, it's probably PEBCAK

EDIT: And, yup, that cleared up the performance issue I was seeing in the cluster too. Nice.

 

[sryan2k1]

*sigh*

Anyone know where we can get a pile of Dell PW116 power supplies for cheap?

 

[Danger Mouse]

sryan2k1[/url]":2opjn6mk]*sigh*

Anyone know where we can get a pile of Dell PW116 power supplies for cheap?

Cheapest looks to be through an Amazon third party seller. You may want to contact them directly to get a discount, since Amazon takes a big cut.

Do a bulk purchase directly and they'd probably come out better on the deal, even after giving you a decent discount.

 

[Black_Obsidian]

So back in the old thread, I made this post about discovering why our backups were so terribly slow. That was about a month ago, and since then the new backup server has been ordered, installed, and tested. The benchmark numbers basically speak for themselves:

Old Busted (HP DL180 G5, E200, 5x1TB 7.2K RAID-5)

New Hotness (Dell R720, H710p, 14x900GB 10K RAID-50, 2 spans of 7 disks)

Both machines idle, and the R720 hasn't even finished background initialization of its array yet.

Still, I think it'll probably be sufficient

 

[whowd]

Black_Obsidian[/url]":17w22hi8]So back in the old thread, I made this post about discovering why our backups were so terribly slow. That was about a month ago, and since then the new backup server has been ordered, installed, and tested. The benchmark numbers basically speak for themselves:
Both machines idle, and the R720 hasn't even finished background initialization of its array yet.

Still, I think it'll probably be sufficient

Looks nice -- So you know, the blank disk in the new system will be at it's fastest now due the physical geometry. You're reading and writing data to the longest outside tracks. As the drives and the RAID volume starts to fill you will hit the interior shorter tracks and it will have a slight hit to your transfer speeds.

w

 

[Mike Bridge]

gradster[/url]":ws6kwnmf]

Mike Bridge[/url]":ws6kwnmf]that i need to update the certificate(s) for vcenter 5.1's Single Sign On component. and it does not look pleasant:

http://kb.vmware.com/selfservice/micros ... Id=2035011
which refers to:
http://kb.vmware.com/selfservice/micros ... Id=2041600
(along with http://kb.vmware.com/selfservice/micros ... Id=2044696 )

Hopefully this can make it easier for you.

http://blogs.vmware.com/vsphere/2013/04 ... lable.html

yeah, the download is the 2nd link in my post. wasn't as bad as i feared, just time consuming. once the cert error was gone from the log, it started throwing a kerberos error, so i left/rejoined the domain with it, and voila, it works again!

 

[SandyTech]

TIL that apparently logmein rescue's technician console doesn't support OSX. I was looking for an excuse to put more RAM and an SSD in my MacBook anyhow.

 

[Danger Mouse]

...that somewhere between the 4.2.1 and 5.0.1 NX-OS code for the Cisco MDS9124, they fixed WTF bug it was that stuck the mgmt port at half duplex.

The code transfer first was going at 70kb/sec. I didn't check what it was like after the upgrade, but it felt nearly instantaneous in comparison.

---

In other news, had a VERY nice interview with the Security Director at HUGE MEDIA CONGLOMERATE. It was promised that the role would not be a dead end.

The Directory readily acknowledged that I have a very wide (and useful) skillset and pointed out that they like to keep talent internal to the company, even if not in the same division. The current guy is moving on to one of their data centers to do virtualization related duties.

Awesomesauce.

We'll see if they offer me the job or not.

 

[WingMan]

Technarch[/url]":22355j1j]

WingMan[/url]":22355j1j]

Altiris and work properly are oxymorons. It's gotten worse since Symantec took it over.

++ Pre-Symantec, we tried setting up Altiris in the lab as a trial. It just wouldn't work. We called the Altiris rep and said, "It's not working." Rep said, "Okay." No attempt to get us support or anything, it was like he knew it was doomed to fail.

Supposedly there are working Altiris installs in the city, but I haven't actually laid eyes on them.

We actually had a working Altiris installation last summer. Had a jr. level systems admin work on it full time. Worked great until the helpdesk started using it. They somehow managed to FUBAR the system in 2 weeks to the point were it needed another complete rebuild. Our old Altiris server from before the Symantec acquisition still works but it's long in the tooth on a G4 DL380. I just swapped in a new MB for it because the memory controller went tits up.

 

[Danger Mouse]

WingMan[/url]":3alvcfoy]

Technarch[/url]":3alvcfoy]

WingMan[/url]":3alvcfoy]

Altiris and work properly are oxymorons. It's gotten worse since Symantec took it over.

++ Pre-Symantec, we tried setting up Altiris in the lab as a trial. It just wouldn't work. We called the Altiris rep and said, "It's not working." Rep said, "Okay." No attempt to get us support or anything, it was like he knew it was doomed to fail.

Supposedly there are working Altiris installs in the city, but I haven't actually laid eyes on them.

We actually had a working Altiris installation last summer. Had a jr. level systems admin work on it full time. Worked great until the helpdesk started using it. They somehow managed to FUBAR the system in 2 weeks to the point were it needed another complete rebuild. Our old Altiris server from before the Symantec acquisition still works but it's long in the tooth on a G4 DL380. I just swapped in a new MB for it because the memory controller went tits up.

You could move the images off to an iscsi target presented as a junction within the Altiris install folder after imaging the system onto newer hardware

Alternately, reinstall onto newer hardware with same host name and IP and then copy over the db and program folder. For best results. Detach the DB cleanly.

Also note the biggest issue we have is with the DB.

 

[KD5MDK]

We had a working Altiris setup at my old job about 5 years ago. Well, working in the sense that I think 90% of the clients were functioning at any given time.

 

[neo7911]

Discovered that Microsoft support is completely worthless when it comes to troubleshooting windows cluster services not releasing persistent scsi-3 reservations. Was working the windows admin and dba encapsulating some Dmx-4 metas behind our vplex.

Ended up failing back to direct Dmx connect and shut down again and it finally released the lock on the one meta.

Also learned how to look up in the Dmx and CX/VNX the locks.

 

[El Guapo]

TIL what a strange feeling it is to vmotion your vCenter VM from within your vCenter VM.

"Yo dawg...."

 

[Darkseid]

10krpm 9gbg LVD scsi drives still hurt like hell when chucklefucks, doing drive swaps, oops butterfingers them rather than handing them down to you.

Once I can forgive, three is 'drop one more and im going to beat you to death with the this backplane'

 

[afidel]

Darkseid[/url]":2bzt1ply]10krpm 9gbg LVD scsi drives still hurt like hell when chucklefucks, doing drive swaps, oops butterfingers them rather than handing them down to you.

Once I can forgive, three is 'drop one more and im going to beat you to death with the this backplane'

Do I even want to know why you're wasting power on 9GB drives?

 

[Paladin]

LVD SCSI drives? On the toes? Those things are tanks, not fun!
I love listening to the sound of them powering up in a stagger start. Like a deck of an air craft carrier full of jet turbines spinning up.

 

[Big Wooly Mammoth]

No, you don't want to know.

 

[Cookie.Monster]

That long ago Microsoft set up a registry key in a beta of Windows 95. They realized this was the wrong way to do things, removed documentation for the RTM, provided a workaround, but left the key in place.

Which many a developer discovered and implemented.

Including someone developing Java.

These don't play well with mandatory profiles. The registry data points to paths in the Mandatory Profile itself (in our case, Cusers\user) no matter who logs in.

All the while the vendor is trying to give the run-around - trying to persuade us to modify privileges, pointing fingers at various layers of the delivery (Is Citrix configured right? I'm not sure if we support a central java cache? Did you check the permissions?).

Ahh well. One XenApp application that includes a script updating "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Shel
l Folders\Desktop" to include the actual userprofile, and things work without issue.

<3 to procmon and developers.

 

[macker0407]

That MS's SQL licensing for VMs is still somewhat irritatingly complex and seems to imply that you have play stupid affinity games with VMs and CPUs if you use the Core model with Standard to remain completely within compliance.

My only hope now is that I can get away with using the CAL model instead, without having to engage in jiggery pokery.

 

[afidel]

macker0407[/url]":3563j25o]That MS's SQL licensing for VMs is still somewhat irritatingly complex and seems to imply that you have play stupid affinity games with VMs and CPUs if you use the Core model with Standard to remain completely within compliance.

My only hope now is that I can get away with using the CAL model instead, without having to engage in jiggery pokery.

Huh? For 2012 you just pay per vCPU as long as you have SA.

 

[macker0407]

That depends entirely on how you parse this, under "Additional licenses are required when:":

Multiple hardware threads are supporting a single virtual core simultaneously. (A core license allows a single v-core to be supported by a single hardware thread.)

Broadest interpretation is that all hardware threads in a host server are supporting all vCPU's simultaneously as the scheduler could put the VM's timeslice on any available CPU, unless you use affinity rules to limit things. The narrowest interpretation implies you only need to licence the vCPUs, based on current hypervisor designs(i.e. a VM cpu instruction isn't further broken into smaller pieces and executed on multiple cores at the same time).

SA just solves the vMotion problem.

 

[wseaton]

I learned that a homeless shelter with no more than 7 concurrent administrative desktop users is a perfectly acceptable location for a Cisco UCS and ESX SANs.

 

[Digitlman]

wseaton[/url]":2j0wr5l2]I learned that a homeless shelter with no more than 7 concurrent administrative desktop users is a perfectly acceptable location for a Cisco UCS and ESX SANs.

Go on.....

 

[Widger]

wseaton[/url]":hcl0qjpj]I learned that a homeless shelter with no more than 7 concurrent administrative desktop users is a perfectly acceptable location for a Cisco UCS and ESX SANs.

I can't wait to hear the rest of this.....

 

[ranald]

wseaton[/url]":3f5k45hs]I learned that a homeless shelter with no more than 7 concurrent administrative desktop users is a perfectly acceptable location for a Cisco UCS and ESX SANs.

Grant money that needed to be spent, no doubt. Or someone's budget would be cut the following year.

 

[Horatio]

I learned that storage spaces on passthrough disks does not work very well. Painfully, it works until the first host reboot. Then you wonder why your VM won't start and can't find its disks.

 

[Rick25]

That I really dislike the day after patch day where anything that's wrong with the app is suddenly blamed on the OS....