Posts made by PhlipElder

@dave247 said in SAS SSD vs SAS HDD in a RAID 10?:

I'm planning the build on a new server. I originally intended on putting 8 x "900GB 15K RPM SAS 12Gbps 512e" drives into a RAID 10 config using an H740P adapter, but then I saw that there are quite a few options for SAS SSD. I haven't really learned too much about the differences of putting SSD in RAID and how it compares to HDD in RAID, so I wanted to see if anyone here (Scott) had any input on the matter.

Example: Would it be worth putting, say, 6 x "1.6TB SSD SAS Mix Use 12Gbps 512e" drives into a RAID 10 instead? Is there a better approach with SSD in RAID?

RAID 6 is the way to go. We lost a server after replacing a drive and it's RAID 10 pair decided to drop out about 5 to 10 minutes into a rebuild.

In our comparison testing 8x 10K SAS drives in RAID 6 has a mean throughput of 800MiB/Second and about 250-450 IOPS per disk depending on the storage stack configuration.

SAS SSD would be anywhere from 25K IOPS per disk to 55K-75K IOPS per disk depending on whether read intensive, mixed use, or write intensive. There are some good deals out there on HGST SSDs (our preferred SAS SSD vendor).

@irj said in Edge (Sync) Favourites Toasted:

That is why I never use Edge! Way worse than IE in my opinion.... It may be faster, but it has so many issues it's unbelievable.

Up until about a week ago with the last round of updates Edge was behaving really well. In fact, the only time I'd had to pull out Firefox was in troubleshooting what is looking to be a ShareBand problem (we have a bonded vDSL 25/5 pair) that caused larger downloads to stall and die.

@dafyre said in Edge (Sync) Favourites Toasted:

At least you were able to recover most of your bookmarks!

Edit: I have also started using NextCloud and the Bookmarks app. No syncing required to get the Bookmarks on all your devices. There are even apps that will let you add bookmarks from your mobile devices too.

Unfortunately not. I lost the week's worth of bookmark/favourites that were a part of the project.

There's a small silver lining: All code and PowerShell tweaks/steps that got modified in VSCode had a direct HTTP/HTTPS link commented into the line so it's not a total loss.

I did a write-up of what happened here: http://blog.mpecsinc.ca/2018/09/warning-edge-sync-ate-all-favourites.html

Make sure to back up the Favourites in Edge after a solid day of adding links.

I fly in later on Saturday and out the following Sunday.

I finished the first three books of Barsoom Series Complete Collection - 11 Books (13 Novels) of John Carter - Fully Illustrated by Edgar Rice Burroughs.

After watching the John Carter movie I got curious about the original source and picked it up on Kindle.

It took a bit to get used to ERB's style given the era it was written in but it was non-stop gripping.

When I have a bit more time I'll finish the series as I got distracted by Conrad Black's Donald J. Trump biography that my foster dad passed along to me.

IT Team gets together and creates the hold from hell.

https://www.theregister.co.uk/2016/04/29/it_helpdesk_creates_oh_hold_hell/

@pete-s said in GeekBench Results for Cloud Servers:

disk

One cannot test for Disk numbers in any real way at least not reliably.

All of the hosting environments we set up have a set of storage QoS policies that are configured according to their customer facing plans. Azure does this and I'm sure that AWS and others also do the same.

EDIT: Better said: One cannot test for in-guest Disk numbers …

@scottalanmiller said in GeekBench Results for Cloud Servers:

AMD Epyc is more recent. I'd like to see more of that, too. But you can't just move from Intel to AMD in a single environment. Very few providers have moved to Epyc because one reason or another, mostly from established norms, I would imagine.

Intel's partner eco-system is one of the best out there IMNSHO. We get great support from them with advanced warranty replacement for any product found to be defective.

When we ventured into AMD a while back to see what was up their eco-system was too fragmented with too much room for willy wagging between hardware vendors whose components were in the servers we were building.

Most, if not all, of Microsoft's A Series VMs back in the day were built on AMD Opteron. When running in-guest performance tests for CPU on the equivalent Intel and AMD instance (Core/vRAM) the Intel absolutely killed AMD for performance.

I'd like to see the same side-by-side comparisons of the current AMD EPYC and Intel Scalable. AMD most certainly has the PCIe Gen3 limitations somewhat mitigated by adding all of those lanes, but that becomes a somewhat moot point if the CPU is not as efficient as the equivalent Intel product.

@storageninja said in Looking for firmware:

@nerdydad said in Looking for firmware:

I contacted VMware support and am thinking that they are full of crap. I'm on driver version 12 and supposedly need to be on version 16. I'm about ready to bang my head against the desk.
Thank you for taking my phone call today. We discovered that the ESXi host connecting to the storage with external HBA "Mass storage controller: Avago (LSI Logic) Dell 12Gbps SAS HBA external" The driver for that device is "lsi-msgpt3 version 12.00.02.00-11vmw" on the ESXi host version 6.5 U1. The ESXi 6.5 U2 host that is not able to communicate with storage have driver version "16.00.01.00-1vmw.650.2.50.8294253" I presume that this issue is related to the firmware version on the PCI device that we try to communicate trough. Please contact your hardware vendor in order to get correct combination and update it accordingly to their recommendation. I will keep this case open and wait for your update as agreed.

Why do you think they are full of crap. Ancient drivers can do really weird stuff with firmware versions they were never tested with. The External HBA in general is a crazy rarely used device and historically has fun bugs. If you don't like what they said call Dell...
I promise you LSI/Broadcom will ignore your support requests if it's a Dell part.

Concur. Most vendors are pretty good about getting compatibility matrices in their ReadMe.TXT files … most.

Whether it's hardware and the OS to go on it, firmware and PCB versions, firmware and driver versions, and other such combinations one needs to be mindful of their interactions or failures to interact. We have a parts bin full of broken promises in that department as we build and test all solution sets before deploying.

EDIT: I didn't complete my thought: Old and new mix do they not.

We run NETGEAR XS712T and XS716T at the entry-level for full RJ45 10GbE and other than a few firmware glitches they've been flawless as the fabric between storage (SOFS) and compute (Hyper-V).

For Gigabit the SG500x series Cisco switches (the Cisco/Linksys child) have been solid whether standalone or stacked.

We have a number of environments with both sets of switches, a pair of each, and they've been running together for years now with very little issue.

Need the service tag to pick up the right driver and firmware.

We've seen some really nasty bugs patched by some of the latest SAS HBA firmware out of Dell (they cook their own).

We deploy converged solution sets on their stuff (2+ nodes + JBOD(s)).

@scottalanmiller BTDT

Back in the day the general sentiment was that RBL "services" were no more than extortion rackets. IMNSHO, that has not changed much.

With the advent of SPF, DMARC, and DKIM their relevance will become a lot smaller which is a happy.

@wirestyle22 said in Nested hypervisors?:

Maybe

Guest Clusters have been around for a while on the Hyper-V platform. Backing up the shared storage being used by the guest clusters is another matter that has not been solved as of yet. At least, not cleanly.

There are some business cases for guest clusters such as Exchange or SQL whose teams require their own resilience measures be in place in order for the setup to be "supported" in the event something goes wrong.

EDIT: Meh … low blood sugar … need to eat as it's been a busy day already. I totally missed "guest cluster" versus "nested hypervisor".

@obsolesce said in Nested hypervisors?:

@dustinb3403 said in Nested hypervisors?:

@wirestyle22 said in Nested hypervisors?:

@dustinb3403 said in Nested hypervisors?:

@obsolesce said in Nested hypervisors?:

@wirestyle22 said in Nested hypervisors?:

@dustinb3403 said in Nested hypervisors?:

@wirestyle22 said in Nested hypervisors?:

@emad-r said in Nested hypervisors?:

Yup go figure for value, Virtualbox has no intention of doing this amazing feature since 2011

Maybe I'm missing something but why in the world would I ever want to use nested hypervisors? Vendor requirements?

If a vendor is stating that they only support a specific guest OS on a specific Hypervisor they had better be supporting the entire stack and not just an application that is on the guest. . .

What scenario is a nested hypervisor useful in any way?

The only realistic "production" usage for nesting, would be if you for example want to give a Dev his/her own hypervisor to cycle through VMs... or some similar situation where you can't dedicate hardware to.

Even in this case, would nesting be required?

Why not do permission based limitations so you can provide a Dev with access to create/destroy as many VM's as he/she needs within the constraints of your pool or resource limits?

I mean, isn't it likely to be his own host completely? I wouldn't let someone manage a host I'm responsible for

Well. . . no

Just as an example, with XenServer (and XO) you can create users and give them access to a specific pool or set amount of resources on any server in the pool, and to what guests they could affect.

So this would allow the user to do their job without the need for additional hardware or nesting. Unless their job was to develop on a specific hypervisor.

To get that kind of control wiht Hyper-V , you need SCVMM.

We don't have that anymore.

On top of that, he was good with KVM. So I had created a nested KVM host on Hyper-V which was great for him for a while. But he was coming from VirtualBox, and didn't have time to convert all the stuff and whatever else that was involved.

AzMan (Authorization Manager) was the method we used to delimit host access permissions up until 2012 R2. It's been deprecated which is sad as it actually worked quite well.

On 2016 going forward, SCVMM and I think Windows Admin Center may also have tiers but we've not really had any time to invest in WAC as of yet.

@scottalanmiller said in What Are You Currently Reading Outside of Tech:

Romania uses directional LEDs to reduce pollution and energy consumption. Works pretty well. About the same light on the roads, far less in the skies or in your faces, way less power needed.

Part of the irony in that is any fired power plant such as coal and natural gas can only spool down so far before they can no longer spool up without a start-up procedure.

Our coal and natural gas fired plants have a huge rod that goes into the ground for the surplus energy.

@jaredbusch said in What Are You Currently Reading Outside of Tech:

@phlipelder said in What Are You Currently Reading Outside of Tech:

I wish we were more like cities in the US where residential areas only get key intersection lights.

This exists no where in the US that I have lived.

The only reason lights are limited is because the local government ran out of money and has to wait a few yeas for more lights.

My Dad is in St. Petersburg, FL. They limit the number of street lights there in residential areas. My assumption my bad.

@dustinb3403 said in What Are You Currently Reading Outside of Tech:

@phlipelder Linux has a several really amazing star tracking and charting apps if you haven't looked em up.

We have a relatively low pitch roof (~4/12 or 5/12) that is L shaped. One of our dream To Do items is to set up a platform on the garage side of the roof and put a good sized telescope up there with a full remote configuration. It's been on the To Do list for a couple of years now as it got put on hold because Edmonton swapped all of their street lights for LED over the last 24 months.

Prior we were outside the light envelope of the city. Now, we've lost some of our night sky due to the amount of light "pollution" they throw off. I wish we were more like cities in the US where residential areas only get key intersection lights. That would greatly reduce the amount of light pollution and give city dwellers a better sleep.

As long as there are no clouds over Edmonton we can still pull it off though. The Little Dipper is fairly easy to pick out in the night sky which is the boundary for good night sky viewing. The Milky Way is amazing and the binary in the Big Dipper is neat to pick-out with the small telescope we currently have.

We have a HP colour 11x17 printer at home (we home school) so I print the following out regularly:

http://www.skymaps.com/

An amazing interactive sky map:

http://www.sky-map.org/

I love the stars and can spend hours at night watching them.

A super cool moment was teaching each of my kids to triangulate to hunt satellites and the Space Station (this one is a big lightbulb going over us).

We set up our Intel RMM and Dell iDRAC Enterprise KVM/IP setups on the internal LAN with a static IP address. Self-issued SSL is fine for this.

Older RMM/iDRAC units may need a legacy Win7 VM with IE9 around for those moments when they need to be managed (we keep one turned off but around for this exact reason). This is especially true with the certificate structure changes that have come through recently. Modern browsers refuse to connect to legacy web management consoles.

Rules are set up on the edge to allow both inbound and outbound packets for their services to our office IP address. VPN is another method to gain access if the edge supports it.

Username and password are set up with both settings being custom with info kept in KeePass here. Make sure to change the Dell default setting! Intel gets set up in the BIOS before it allows site authentication and access.

Cost wise Dell is $300 to $450 here in Canada to license while the Intel RMM module is sub $150.

A blog post on what we do: http://blog.mpecsinc.ca/2017/06/disaster-preparedness-kvmip-usb-flash.html

We don't do SuperMicro.