Topics created by Pete.S

@Pete-S said in Windows: Set screen to blank when locked?: Sorry, I should have explained more clearly what I'm try to accomplish. I want to screen to blank, which is what causes the screen to go into power saving mode. Not just setting it to a black color. Blanking is tech speak for when the graphics adapter would turn of either one or both of the horizontal and vertical sync signals and thereby communicate to the display that the screen is not in use and it can go into one of it's power saving modes. Lock windows computer =>> screen goes into power save mode right away. Computer don't go into sleep until it does so by it's normal settings. So then this might apply to this

@black3dynamite said in Pros and cons running local repository (for packages like deb, rpm etc)?: @Pete-S said in Pros and cons running local repository (for packages like deb, rpm etc)?: Cons would be storage space - amd64 repository for debian stable is around 60 GB. According to the information in this link, which is updated daily, amd64 is 330 GB. https://www.debian.org/mirror/size It's because that includes everything like testing, unstable, experimental and what not. Also source code. So it's a couple of different versions of the OS. The amd64 stable main contrib non-free are 60.1 GB today - as of right now...

Short of getting advisories from vendors, just putting something on the calendar is probably best.

@Pete-S said in Which hosts belong in what pool when running local storage?: I'll just want to confirm that I will go with one pool for production and one pool for development and benchmarking. Citrix recommendation for Xenserver is to have production & development hosts in separate pools and it makes sense just from avoiding avoiding human mistakes as well. Also all the hosts in the pool share the same network config so it also makes sense to have development in it's own pool in case one want to test something out. Ideally, I'd have one set of switches for each pool as well, but I'll have to settle for different vlans for now - for budgetary reasons. Sounds like a reasonable plan.

@Pete-S said in Geekbench observations: @dafyre said in Geekbench observations: @Pete-S said in Geekbench observations: The relationship between the single-core and multi-core score should be about 80% of theoretical max on the multi-core score. So if single core score is 3000 and you have 4 vCPUs then multi-core score should be 80% of 3000 x 4 cores = 9600. If the host is under heavy load the multi-core score will go lower and lower. I think you are on the right track. This is largely in part due to how the underlying Hypervisor handles multi-core VMs. The way I understand it, is that in a multi-core VM, the Hypervisor has to wait for that number of cores to be ready to process before it signals to the VM that it can keep running. IE: In your example, a 4 core VM, the underlying hypervisor will have to wait to have 4 cores waiting for work before it will tell the VM that it's cores are available. I've read that before but I think it is some old feature of very old hypervisors called strict co-scheduling. It's not used anymore. Nowadays basically every hypervisor has their scheduler that puts vCPU on real pCPUs according to the time share principle. So every vCPU get's a piece of the pie. But it has to account for hyperthreading, more than one CPU socket (NUMA), power saving, VM priority and other things. The underlying principle is though that all VMs and their vCPUs should get their fair share of CPU time. Some hypervisors have different scheduler algorithms so you can pick other ways of scheduling that might be more optimized for your workload. Depends, SMP doesn't really allow for that, all cores have to be in lock step. Only is AMP is supported can the hypervisor do that. It requires the hypervisor and system above it together to do non-SMP processing.

The only scenarios I've seen are: PPPoE / DSL Jumbo Frames VPNs

@Pete-S click the menu (three stacked dots) by the post that had the answer so others know which solution worked best for you.

@EddieJennings said in Typical virtualization host server config?: @Pete-S said in Typical virtualization host server config?: We've calculated guesstimated 8GB RAM and 50 GB SSD storage per VM on average. How did you come up with your guesstimate? Do you have some workloads in production right now that you'd be virtualizing with whatever host your configure and buy? Yes, actually we have the hardware now but it hasn't been configured completely yet. We are still figuring things out. The workloads are running on physical servers (Windows) and it's latency that is our primary concern, not capacity. We are moving the applications to linux and making them multi-tenant and horizontally scalable at the same time. So that's why we are guesstimating based on the config of the physical servers.

Sounds like a good idea guys. I will do it that way.

I'd be interested in hearing what you try out and how you are using the solution.

Went to a store the other day where they had a large number of different LEDs on display. Checked them out and all Philips LEDs were flicker free but the noname ones were a mixed bag. Some were flicker free while others were horrible and some in between.

@wrx7m said in Load balancer inside firewall or not...: Wasn't pfsense in production, I guess I would say, "frowned upon"? PfSense is awesome. Avoid turning into a UTM so it can stay stable.

@Pete-S Not me. I have no need for top of rack quality gear at any clients. I generally avoid Dell network hardware just because HP usually has a comparable solution.

https://forum.xda-developers.com/android/apps-games/galaxy-playstore-alternative-t3739733

@tonyshowoff said in Thicket files in Windows - how are they "linked"?: @Dashrender said in Thicket files in Windows - how are they "linked"?: I recall seeing this ages ago. It happened when I would save a website onto my computer. I found that the directory had copies of the images (and perhaps other files) that the HTM file referenced. @Pete-S Indeed this has been true since at least Windows 98 when IE had the feature (maybe still does) "Read web page offline" and also as a part of the Save... feature, and other browsers copied that same implementation when saving the assets. In the end it doesn't matter where it came from, how it was created, etc nor whether 9x or NT. If there's a .htm(l) file and a directory with _files in the name they are treated as linked in Explorer, but not in CMD (IIRC). Thanks. I made the same observation - that it is the name that link the directory and file together and it is only in Explorer it is treated as a link.

Sweet.

@stacksofplates said in Documenting rack, servers, drives, CPU, RAM etc: @dbeato said in Documenting rack, servers, drives, CPU, RAM etc: I use https://www.draw.io/ and I know a tool that @stacksofplates recommended for Cloud Diagrams. Draw.io recently implemented one that looks almost identical to the one I was using so I've been playing with that. The other was cloudcraft.co. For rack diagrams, phpIPAM isnt too bad. And you get the benefit of IPAM software with it. Draw.io is in NextCloud too.

I should have noted that grep likely wouldn't like complicated RE patterns. Sed, however, supports the full gamut and is installed by default on pretty much every linux variant; not sure if pcregrep is. It seems everyone likes the link I posted for the RE builder so I'll paste two others that I use: https://regex101.com/ http://leaverou.github.io/regexplained/ Cheers!

@scottalanmiller said in PVLAN (private VLAN) in the switch - are you using it?: PVLAN, or Port Isolation as I think most of us know it, is one of the better uses of VLAN tech. The idea is for extreme environments (not really SMB generally) when normal security measures are not enough, that you make an individual VLAN for every single device on the network so that you control via central firewall a second layer of access for every single port that there is. There are certainly legit cases for this. And I've worked for one of those places. But it's super rare. It is a lot of work, requires gear that supports it, and adds a lot of complication that you have to consider. It also adds a good deal of security. In the SMB, most places have over the top security already and zero day threats rarely threaten OS level firewalls. So PVLAN, while legit, rarely has appreciable value to an SMB. But when you need that "second firewall per device", then yes, it's definitely the way to go. Makes sense, but I'm thinking it doesn't have to be that much more work if you can apply automation to switch management as well. I think you can do port isolation on the virtual switches in VM hosts in the same way as the physical ones. I understand that at least VMware has had it for a long time so assume other have it now as well.

Cheap meaning low cost, of course.

@black3dynamite said in Automation with Ansible, Salt etc - at what point?: @obsolesce said in Automation with Ansible, Salt etc - at what point?: @black3dynamite said in Automation with Ansible, Salt etc - at what point?: @scottalanmiller said in Automation with Ansible, Salt etc - at what point?: @black3dynamite said in Automation with Ansible, Salt etc - at what point?: @obsolesce said in Automation with Ansible, Salt etc - at what point?: @black3dynamite said in Automation with Ansible, Salt etc - at what point?: @scottalanmiller said in Automation with Ansible, Salt etc - at what point?: @pete-s said in Automation with Ansible, Salt etc - at what point?: Ansible seems to be the least complicated to get started with so I guess that'll be as good as anything. Syntactically yes. Salt has the simpler architecture, because it is clients reaching the server, not the server reaching the clients. That’s the main thing I like about salt. But damn, if the minion service is hosed for whatever reason can be a real pain. SaltStack can do agentless as well, like Ansible. Ansible uses winrm to manage Windows. Can Salt do the same? Because Salt agentless uses SSH, so I would need to setup ssh server on Windows. https://docs.saltstack.com/en/getstarted/ssh/index.html Why would you want to do that, though? The agent is the key reason to be on Salt in the first place. I'm all good with using the agent. But until I figured out the problem I'm having the agent on my Windows machines, Ansible will be used. I've got the agent deployed across 700 win7, Win10, win server, and Hyper-V servers at work. All working, installed via chocolatey. What is the issue you are having? Edit: 50-100 of those 700 are Linux. Its probably something stupid on my part but It's only happening on some of my Windows 10 1803 machines. They are installed via chocolatey too. The service gets stuck in a paused state. It is working great on the other Windows 10, 7, servers, Hyper-V and Linux. I think that means it has no contact with the salt master.

The place that I see them most is in DCs that allow questionable public access rather than having dedicated, screened DC workers only getting access.

@donahue said in My first computer: @travisdh1 said in My first computer: @scottalanmiller said in My first computer: @travisdh1 said in My first computer: @scottalanmiller said in My first computer: @pete-s said in My first computer: Which reminds me of IBM's PS/2. What a train wreck that thing was. Guess what made up the entire computer lab in my high school... Yep IBM PS/2. They actually network booted. That's crazy. that must be all that IBM ever managed to sell. My theory is that some company bought a whole bunch of them, realized they were terrible, and donated them to the school for a tax write off. They had all apple IIe's at my computer labs growing up. I was always told it was because apple basically gave them to schools for near free, maybe at cost or something, and that it was part of the strategy to get people on one particular platform at an early age. It was always inconvenient because the files were never compatible between our home PC and the ones at the schools. That and the Apple ][ were just awful. I still have the one from my elementary school.

@fateknollogee said in Small colo infrastructure - rack layout feedback: @pete-s The fs.com website is what every decent supplier should strive for. Quantity in stock is shown clearly, no guessing needed! They tell you when the stuff will ship!! Yes, that's the way it should be. I saw that they also do custom orders so if I need 200 purple power cables that are exactly 5.5 ft long they can do that. I'm already putting together my first order from them

@travisdh1 said in How secure are databases in general?: They are vulnerable, but mostly because applications use a single authenticated user to access a given database. So most of the vulnerabilities are in the web pages and/or applications using a database. That's application security, not related to the database. Same risk would exist with no database.

@dave247 said in Questions on redundant switch setup: @pete-s said in Questions on redundant switch setup: @dave247 said in Questions on redundant switch setup: @pete-s said in Questions on redundant switch setup: @jaredbusch said in Questions on redundant switch setup: @dave247 said in Questions on redundant switch setup: @pete-s What kind of firewall and switches are you running? One option: if you're switches have stacking, then you can put them in a single stack and then create a port group that spans the two switches and then connect that to your NIC teams on the other end. This guards against switch failure, switchport failure, server NIC port failure, Ethernet cable failure, etc.. This adds a level of complexity that you don't have to deal with when using a simple team. But the plus side is higher bandwidth per connected server. The "switch independant team" what bonding mode is that in linux? Is it mode 1, active/backup policy? You will have to look at your individual network card's drivers and management software with regards to Linux. AKA, read the manual. My guess is that you're running Broadcom NICs and the management software that I've seen/used is called "Broadcom Advanced Control Suite 4" and the "switch independent mode" or team type is called, "Smart Load Balancing and Failover (SLB)". I'm all Intel on the NIC side in this case as Supermicro is predominately intel NICs and they are very well supported both in freebsd and linux. Contrary to Windows, linux actually have bonding of different types in the kernel (a module called bonding). So the drivers don't have to do bonding. oh nice. I have no idea. I haven't done much with Linux lately. Still, I would read the NIC documentation as it pertains to Linux. Looking at Dell switches it seems like Dell N1124 will do the job. It's 24x1G switch with 4x10G for uplinks and stacking and has most of the features of it's bigger brothers in the N2000, N3000 series. Pricing looks very attractive where I'm at (<$400 USD per switch), otherwise it's $1259 in the dell.com store. I've never used Dell switches though.

@stuartjordan said in Freelance websites?: You have two good companies that would be able to help you from this forum: NTG or Bundy Associates, give them a shout. LOL yep I was late

Alright, I took another approach at this. Looking directly at the source for different kernels over at kernel.org I can see that the NVMe driver has SR-IOV support from kernel version 4.8. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/nvme/host/pci.c?h=v4.8#n2015 It's very interesting for virtualized workloads because SR-IOV means you can assign the NVMe disk to different VMs directly and bypass the hypervisor. This is the same for NICs at 10/25/40/100 gigabit. There will be a significant performance drop if you don't use SR-IOV.

@dustinb3403 said in Small colo infrastructure for SaaS: @pete-s said in Small colo infrastructure for SaaS: This is how @dustinb3403 suggested replicating VMs between hosts: While this is accurate, it also misses on the fact that he would still have his NLS server sitting, collecting backups on whatever schedule. Other than that it is accurate. In Scott's proposal you are making the shift from migrating the entire workload (which is essentially instant) to migrating the database only. In his case, the load balancer is the weak link in the chain. Granted these don't fail often but it isn't something you have control over either unless you provide your own for the COLO. 2 haproxy VMs (one per host) and keepalived for failover

@Donahue said in Why you don't need a VPN or not?: @scottalanmiller said in Why you don't need a VPN or not?: @coliver said in Why you don't need a VPN or not?: @Dashrender said in Why you don't need a VPN or not?: But I don't see that working very well for large files - say AutoCAD or even some graphics files. AutoDesk and Dessault both have toolkits to get around SMB limitations. I think most CAD vendors do actually. Yes, CAD is something that basically always does LANless on its own. It has special needs and normally handles them. You don't use NextCloud or anything like it, nor do you use SMB or anything like that. only for a premium, but they are rarely if ever rolled into the base product, at least not fully featured Technically, all fire sharing is a sort of premium somewhere.