TL;DR. Seems quite good overall. Ping from home is roughly 126ms which is basically what you would expect (my ISP directly peers with Lumen so it pretty much goes to Paris and then directly to Oklahoma)
I've also noticed latency to Kansas City is a bit over the place (26ms/30ms). One of my boxes goes to Chicago and then directly to Oklahoma. Not that bad, but roughly 13ms or so increase.
My other box goes directly though Cogent until it hits Lumen basically. (Omaha - Denver - Tulsa)
That one makes us look good, I like it. Time to figure out what's going on with Hivelocity Chicago (and T-Mobile Fiber Denver.)
Okay interesting, so Hivelocity Chicago is through Lumen, that's good. But I have to figure out how to make BGP to do that. I'm sure it'll be really easy. Wait, what's BGP? Ohhh... ohh... oh.
I haven't found a good route to Denver except that one, okay I've hacked the mainframe, I'm in. That one is Lumen as well. That's real good. Alright time to contact Lumen and see if they can make this easier.
@cybertech said:
wait what happened to virtfusion?
@cybertech said:
wait what happened to virtfusion?
I think VirtFusion was a "maybe" experiment
It's not for these Ryzen servers. I mean I suppose it could be... I guess I never thought of it that way. I'd just need to trick WHMCS into using it for one location as I don't really want to make an entire OKC duplicate for each plan. For another product line on different servers, it made more sense. Or we could have sub-versions for Ryzen nodes.
Orrr.... I could finally get one of the other nodes up, it's crazy how backwards all our plans ended up going, these Ryzen were supposed to probably be last. Same with IPv6.
@VirMach said: Worst route I've seen is probably 26ms to Kansas City, and it's Cogent, and geolocations for the hops are Michigan, Nebraska, Texas, Colorado. But based on Cogent's map, it'd be slightly off but I think it's actually somehow jumping up to Nebraska, Denver, to Dallas and back up.
To KC 24.756 ms
1 _gateway (193.58.147.1) 2.695 ms 2.667 ms 2.648 ms
2 23.136.108.50 (23.136.108.50) 0.384 ms 0.364 ms 0.421 ms
3 1-1-c17-2.ear2.Tulsa2.Level3.net (4.43.131.85) 3.003 ms 2.982 ms 3.006 ms
4 * * *
5 * * 1299-3356-dal.sp.lumen.tech (4.68.74.166) 8.541 ms
6 * * dls-bb2-link.ip.twelve99.net (62.115.138.64) 8.520 ms
7 kanc-bb2-link.ip.twelve99.net (62.115.139.188) 13.507 ms 13.487 ms 13.311 ms
8 kanc-b3-link.ip.twelve99.net (62.115.113.81) 13.440 ms 13.572 ms 13.500 ms
9 nocix-ic-372038.ip.twelve99-cust.net (80.239.132.187) 13.077 ms 13.089 ms nocix-ic-372039.ip.twelve99-cust.net (62.115.147.181) 13.060 ms
10 100ge-0-38.edge-1.clay.mci.us.as32097.net (69.30.209.201) 13.417 ms 13.559 ms 100ge-103.dist-1.101.swift.mci.us.as32097.net (69.30.209.225) 24.634 ms
11 100ge-53-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.200) 12.991 ms 69.197.140.1 (69.197.140.1) 24.756 ms 24.771 ms
From KC 19.834 ms
1 _gateway (69.197.140.1) 0.167 ms 0.252 ms 0.244 ms
2 100ge-55-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.224) 0.199 ms 0.213 ms 0.202 ms
3 kanc-b3-link.ip.twelve99.net (80.239.132.186) 0.598 ms 0.606 ms *
4 be5627.ccr22.mci01.atlas.cogentco.com (154.54.90.249) 1.370 ms 1.370 ms 1.611 ms
5 be5590.ccr31.oma02.atlas.cogentco.com (154.54.47.38) 4.576 ms * 4.843 ms
6 ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.317 ms be3802.ccr21.den01.atlas.cogentco.com (154.54.165.77) 15.064 ms ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.300 ms
7 be2766.ccr41.ord03.atlas.cogentco.com (154.54.46.178) 12.946 ms be2401.rcr21.b006467-6.den01.atlas.cogentco.com (154.54.88.42) 15.151 ms VIRTUAL-MAC.ear2.Tulsa2.Level3.net (4.43.131.86) 13.640 ms
8 orange.lax05.atlas.cogentco.com (154.54.10.2) 14.464 ms 14.429 ms 14.677 ms
9 193.58.147.1 (193.58.147.1) 19.861 ms 19.854 ms 19.834 ms
To Dallas 8.822 ms
1 _gateway (193.58.147.1) 4.330 ms 4.318 ms 4.314 ms
2 23.136.108.50 (23.136.108.50) 0.370 ms 0.394 ms 0.388 ms
3 1-1-c17-2.ear2.Tulsa2.Level3.net (4.43.131.85) 3.024 ms 3.016 ms 3.063 ms
4 ae1.3515.edge1.Dallas2.net.lumen.tech (4.69.209.114) 8.962 ms ae1.3509.edge2.Dallas2.net.lumen.tech (4.69.206.165) 8.244 ms 8.240 ms
5 * * *
6 be2763.ccr31.dfw01.atlas.cogentco.com (154.54.28.73) 8.877 ms 8.901 ms 8.941 ms
7 be8551.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.54.92.105) 9.382 ms be8553.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.54.95.29) 9.529 ms 9.632 ms
8 be4693.nr61.b059724-0.dfw01.atlas.cogentco.com (154.24.5.166) 9.591 ms 9.592 ms 9.465 ms
9 38.32.15.66 (38.32.15.66) 8.707 ms 15.827 ms 15.815 ms
10 64.44.158.1 (64.44.158.1) 8.882 ms 8.826 ms 8.822 ms
From Dallas 16.887 ms
1 1-148-44-64.reverse-dns (64.44.148.1) 0.172 ms 0.159 ms 0.176 ms
2 * * *
3 be4693.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.24.5.165) 1.055 ms 1.170 ms 1.275 ms
4 be8553.ccr32.dfw01.atlas.cogentco.com (154.54.95.30) 1.249 ms 1.277 ms be8551.ccr31.dfw01.atlas.cogentco.com (154.54.92.106) 1.274 ms
5 be2763.ccr41.dfw03.atlas.cogentco.com (154.54.28.74) 1.242 ms be2764.ccr41.dfw03.atlas.cogentco.com (154.54.47.214) 1.144 ms 1.303 ms
6 lumen.dfw03.atlas.cogentco.com (154.54.10.18) 8.863 ms 1.620 ms 6.580 ms
7 ae2.13.ear2.Tulsa2.net.lumen.tech (4.69.136.210) 6.095 ms 6.112 ms 6.034 ms
8 VIRTUAL-MAC.ear2.Tulsa2.Level3.net (4.43.131.86) 8.885 ms 8.889 ms 8.900 ms
9 193.58.147.1 (193.58.147.1) 16.932 ms 16.910 ms 16.887 ms
@VirMach said: Worst route I've seen is probably 26ms to Kansas City, and it's Cogent, and geolocations for the hops are Michigan, Nebraska, Texas, Colorado. But based on Cogent's map, it'd be slightly off but I think it's actually somehow jumping up to Nebraska, Denver, to Dallas and back up.
To KC 24.756 ms
1 _gateway (193.58.147.1) 2.695 ms 2.667 ms 2.648 ms
2 23.136.108.50 (23.136.108.50) 0.384 ms 0.364 ms 0.421 ms
3 1-1-c17-2.ear2.Tulsa2.Level3.net (4.43.131.85) 3.003 ms 2.982 ms 3.006 ms
4 * * *
5 * * 1299-3356-dal.sp.lumen.tech (4.68.74.166) 8.541 ms
6 * * dls-bb2-link.ip.twelve99.net (62.115.138.64) 8.520 ms
7 kanc-bb2-link.ip.twelve99.net (62.115.139.188) 13.507 ms 13.487 ms 13.311 ms
8 kanc-b3-link.ip.twelve99.net (62.115.113.81) 13.440 ms 13.572 ms 13.500 ms
9 nocix-ic-372038.ip.twelve99-cust.net (80.239.132.187) 13.077 ms 13.089 ms nocix-ic-372039.ip.twelve99-cust.net (62.115.147.181) 13.060 ms
10 100ge-0-38.edge-1.clay.mci.us.as32097.net (69.30.209.201) 13.417 ms 13.559 ms 100ge-103.dist-1.101.swift.mci.us.as32097.net (69.30.209.225) 24.634 ms
11 100ge-53-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.200) 12.991 ms 69.197.140.1 (69.197.140.1) 24.756 ms 24.771 ms
From KC 19.834 ms
1 _gateway (69.197.140.1) 0.167 ms 0.252 ms 0.244 ms
2 100ge-55-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.224) 0.199 ms 0.213 ms 0.202 ms
3 kanc-b3-link.ip.twelve99.net (80.239.132.186) 0.598 ms 0.606 ms *
4 be5627.ccr22.mci01.atlas.cogentco.com (154.54.90.249) 1.370 ms 1.370 ms 1.611 ms
5 be5590.ccr31.oma02.atlas.cogentco.com (154.54.47.38) 4.576 ms * 4.843 ms
6 ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.317 ms be3802.ccr21.den01.atlas.cogentco.com (154.54.165.77) 15.064 ms ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.300 ms
7 be2766.ccr41.ord03.atlas.cogentco.com (154.54.46.178) 12.946 ms be2401.rcr21.b006467-6.den01.atlas.cogentco.com (154.54.88.42) 15.151 ms VIRTUAL-MAC.ear2.Tulsa2.Level3.net (4.43.131.86) 13.640 ms
8 orange.lax05.atlas.cogentco.com (154.54.10.2) 14.464 ms 14.429 ms 14.677 ms
9 193.58.147.1 (193.58.147.1) 19.861 ms 19.854 ms 19.834 ms
That's interesting, and confusing. So is that the same routes I described except Cogent going from Omaha --> Denver --> Dallas --> OKC ends up being better than... oh I see it now, they handed it over to Telia from Dallas. And the Cogent one improved because of Telia in that case.
In my comparison, Telia wasn't involved I guess, just Cogent. It would've probably been helpful if I posted it, luckily I have the tab up. Oh nope it's not because of the Telia, it went to "Chicago" instead of "LAX" or.. is it the "Michigan" -- okay whatever I'm done playing Cogent placeholder hostname shuffle because it also geolocates to the that incorrect hostname still and I don't feel like triangulating each IP address.
HOST: ulanqab-alibaba (Jining - Ulanqab is a common naming for datacenters in this region of Inner Mongolia, Jining being a city within Ulanqab league/prefecture)
NextTrace v1.4.0 2025-04-16T01:10:07Z dccc41b
[NextTrace API] preferred API IP - 45.88.193.28 - 133.59ms - 🐠 (Relay) → Misaka.LAX
IP Geo Data Provider: LeoMoeAPI
traceroute to 221.183.166.210, 30 hops max, 52 bytes payload, ICMP mode
1 193.58.147.1 AS25693 Oklahoma City Oklahoma United States IDC
2.21 ms / 9.61 ms / 8.84 ms
2 23.136.108.50 AS25693 Oklahoma City Oklahoma United States
0.44 ms / 0.37 ms / 0.45 ms
3 4.43.131.85 AS3356 Tulsa Oklahoma United States lumen.com
1-1-c17-2.ear2.Tulsa2.Level3.net 3.09 ms / 3.12 ms / 3.13 ms
4 *
5 4.68.75.218 AS3356 Los Angeles California United States lumen.com
58453-3356-lax.sp.lumen.tech 38.96 ms / 38.95 ms / 39.05 ms
6 223.120.6.217 AS58453 [CMI-INT] Los Angeles California United States cmi.chinamobile.com 移动
38.94 ms / 39.12 ms / 39.59 ms
7 223.120.12.214 AS58453 [CMI-INT] Shanghai China CMI-CM-Peer cmi.chinamobile.com 移动
223.64 ms / 236.04 ms / 223.60 ms
8 221.183.55.110 AS9808 [CMNET] Beijing China X-I chinamobileltd.com 移动
222.25 ms / 222.22 ms / 222.26 ms
9 221.183.25.201 AS9808 [CMNET] Beijing China I-C chinamobileltd.com 移动
222.29 ms / 222.27 ms / 222.27 ms
10 *
11 *
12 221.183.166.210 AS9808 [CMNET] Guangzhou Guangdong Province China chinamobileltd.com
227.61 ms / 227.52 ms / 227.54 ms
From NE Scotland:
3 ae50-ner001.msp.as13285.net (78.144.1.33) 13.928 ms 14.179 ms 14.436 ms
4 ae50-scr001-msp.as13285.net (78.144.1.32) 14.737 ms 15.337 ms 15.668 ms
5 ae60-scr101.thw.as13285.net (78.144.1.110) 21.508 ms 22.113 ms 29.658 ms
6 ae54.edge7.london1.level3.net (217.163.47.149) 30.129 ms 23.205 ms 28.596 ms
7 ae5.5.ear2.tulsa2.net.lumen.tech (4.69.227.145) 118.941 ms 119.222 ms 119.411 ms
8 virtual-mac.ear2.tulsa2.level3.net (4.43.131.86) 122.410 ms 122.703 ms 122.947 ms
9 23.136.108.51 (23.136.108.51) 123.682 ms 123.938 ms 131.552 ms
10 193.58.147.xxx (193.58.147.xxx) 123.228 ms 120.955 ms 121.567 ms
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural. It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
Some high quality Lumen (albeit loves to backhaul through DFW) from all tests I've done.
Do we need to play some sort of Saw type game to get that $25 dedi @VirMach ? This would be the perfect 'not just another DC in the tri-state area' for backups and I need one more bare metal server if it's possible. Let me know who / what I need to sacrifice to get in on that deal.
Now this is confusing (for me): my ATL hostname has a MIA node name, yet is now in OKC.
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural. It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
For privacy reasons, I will not record the IPs of visitors (will uniformly use 0.0.0.0 instead), but will record download speeds, upload speeds, and ISP sources (not very accurate).
All data will be deleted at the end of the next month or before the service stops, and this service will stop at the end of the next month or when the bandwidth runs out.
@VirMach
From WHMCS, VNC works as it should. Change host name works.
From SolusVM, reconfigure networking works on template installed VPS.
I did not found any bugs in the control panels so far.
Will try more when reinstall is available.
Network is nice and steady.
However I recall this feature only in Ubuntu waiting for network and you needed to tweak some settings (netplan (c)), but yours is by disk - checking something?
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
answer: the grease needs to warm up to allow the rust to spin on the disc.
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural. It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
@AlwaysSkint said:
Now this is confusing (for me): my ATL hostname has a MIA node name, yet is now in OKC.
All accurate in a 630 mile radius
@imok said:
I want my Virmach servers in Dallas please.
Order them and I can unplug it and drive it down to Dallas for you, in a self-storage unit.
@Kris said: Some high quality Lumen (albeit loves to backhaul through DFW) from all tests I've done.
I've spoken with our "solution engineer" at Lumen.
So Lumen's closest POP is Tulsa even though they have lines going up out of Oklahoma City. I assume in some specific cases this could mean theoretically, one route that looks closer might actually be further away since it has to come in from Tulsa but we've already confirmed for Denver and Chicago that there should be no reason for it to need to backhaul to Dallas.
They're going to speak with a "WAN Architect" to come up with some suggested changes.
@risturiz said:
DC Lat / Lng? So i can add this to NMS map :-)
35.4668473, -97.6492831
@Kris said: Do we need to play some sort of Saw type game to get that $25 dedi @VirMach ? This would be the perfect 'not just another DC in the tri-state area' for backups and I need one more bare metal server if it's possible. Let me know who / what I need to sacrifice to get in on that deal.
I'm trying to modify the BIOS to have some important setting already set how we need it and trying to flash it with a programmer instead. This ends up saving a lot of time but at the same time it could end... not well. Basically the boards are super annoying with how they end up supporting the newer processors, I've already figured that part out, now we just need to replicate it quickly onto many boards. Otherwise I'm stuck putting a processor on, updating, removing, putting in new processor, and between the processor change the setting resets, and once it resets I lose VGA output. So then I have to connect an external GPU, and I don't have any GPUs that can just get enough power from PCIe right now, so I have to also bring in another power supply, reconnect everything, plug in the GPU, connect it through that, then change the setting, then save, then remove the GPU, rewire everything. Per server. And of course all the waiting for booting, the update to complete, any weird complications, etc.
I should've just done 3000 series to be honest. All of this is because someone at ASRock did the timings incorrectly for something for this specific AMD processor and just "assumed" all 5000 series are the same. I kind of agree with them though, AMD shouldn't have had sub-series or whatever you want to call it for processors, just call it something else, don't pretend it's in the same line. Even worse, within the same series they specifically made for being APUs, they went back and disabled the integrated GPU.
For the remainder I may just do 2600 or 3500 (with the 2600 you get full hyperthreading, with 3500 you don't but it performs the same. The 3500 should have better compatibility with the boards though.) These are the ones that will go in the $25 configuration so I'll let @Kris decide.
If that's the plan I go with though that might mean no 5500 version anymore for the $25 configuration and a slight delay on those until new CPUs come in, I might still be able to do a few of them with 2600's.
@taizi said:
question:why everytime start i need to wait 1min30s
Default SolusVM template. I had some idea for fixing it in our own template version but those never actually get finished.
Yesterday I got additional servers fully ready. The KVM switch is also in place. Rails also came in for some servers. The dongles also came in I think for KVM switch. Cox Communications also asked us for the 3rd time if we want single mode or multi-mode fiber so I've answered them again.
I'm probably going to get some other servers ready, and go in to the DC tonight with them and then maybe redo the cable management and organize things a bit now that it's taking shape. This might mean the first unscheduled maintenance so beta testers may go down for a bit. I also need to change the DAC, I had to use one that isn't ideal length. Except I don't even think I have the new one yet so that might be our second unscheduled maintenance.
IP addresses may also change or I might migrate everyone to a different node.
The storage node has an SSD now, but this board also has some weird VGA issue, I might need to update the BIOS or just fix IPMI do it through that, it needs operating system installed.
Epyc nodes, we have two more that are "ready-ish" and another two coming in, with a sixth one being built-ish. These all have finnicky motherboards that won't accept the memory I have, so I'm ordering additional memory for them as well.
One big issue at the facility is that (a) the WiFi doesn't work which isn't uncommon, and (b) my cell reception has gotten to the level of being completely unusable. So I'll try to remember to take down a WiFi router.
Then I'll probably spend the rest of the time today or tomorrow setting up VirtFusion and/or dedicated server provisioning module of sorts, maybe also setting up the controls for the PDUs. But I'd probably say most my time is going to be spent staring at the cabinets and wondering what should go where, like for way too long instead of actually doing it, so inevitability lots of shuffling positions.
The rest of the weekend, likely no sale, but spring cleaning. I already have too many boxes and too many things I took out of shelves, I need to re-organize inventory. I'm still in OKC, my flight got cancelled so for now the trip I was going to take has been cancelled. So if anything goes wrong then I should be able to fix it pretty quickly after my monitoring system alerts me (you guys on this LES thread at the moment.)
It's looking like this right now. Can change, it's already changed so much, I just never updated you guys until I was semi-sure.
The prices are estimates and are the sale prices. What will likely happen for the sale is that I'll build everything out, won't put in specific drives, and they'll be "pre-orders" except not really a pre-order, as in it'll only be done when they're racked, it would just help get an idea on the demand and then let's say 72 hours setup time or we say we failed to deliver it and just refund you.
Let me know if there's any feedback on how these should be modified, if pricing looks good, and so on. We're pretty flexible with them at this stage. All I know is that I tried to put in both NVMe and a RAID controller and while that's technically possible, it would need to be done later with some custom bracket, specific controllers, etc.
@VirMach said:
Oh maybe I should have mentioned this last part:
Anyone who previously had a dedicated server with us and was impacted by a delay in it being switched to Ryzen and you wanted Ryzen, or you were negatively affected by the time we got Biloh'd or the time we got Ernie'd then we would offer you an equivalent Ryzen bare metal at the same price you were paying. Unfortunately, since I know someone may ask, this doesn't mean we can include ECC memory or KVM or IPMI or more IPv4 even though the original technically had those thing(s.) Those are all additional line items and weren't calculated in my ability to make that offer. Also putting another disclaimer, this doesn't apply to any dedicated servers we may have given out for free. Maybe. Let's say all of this requires my approval but I'll be nice.
People can send in their own drives and I'll install it without any additional setup fees (for members of this community, at least for these initial purchases) as long as you do not ever want it back, even if you cancel (I'll make an exception to this if you're sending us a $500 hard drive or something like that and get pre-approved for me to track it so we know it's yours.)
Wait wait, we're doing dedis again? Oh man, I picked the perfect time to check in again, I definitely want to get in on this. My old dedi was one of those insanely unsustainable deals, I think 225/yr for 32gb/240gb, and you went so far above and beyond already with everything that happened through your dedi ordeals. I'd love to snag one of these, especially with the option of sending in my own drives. I've got a pair of 14tb wd reds I bought for a NAS build that never happened, and rather that keep them sitting around gathering dust, this seems like the perfect use for them.
@bakageta said: Wait wait, we're doing dedis again? Oh man, I picked the perfect time to check in again, I definitely want to get in on this. My old dedi was one of those insanely unsustainable deals, I think 225/yr for 32gb/240gb, and you went so far above and beyond already with everything that happened through your dedi ordeals. I'd love to snag one of these, especially with the option of sending in my own drives. I've got a pair of 14tb wd reds I bought for a NAS build that never happened, and rather that keep them sitting around gathering dust, this seems like the perfect use for them.
Yep, we can do that. Just put in a ticket. So your equivalent based on what you said would be:
@VirMach said: B ) $35/mo - Ryzen 5500, 32GB RAM, 4x500GB** SSD (Can optionally upgrade to 4x1TB SSD)
And it'd be $225 a year or whatever you mentioned instead of the $35 a month. If you want to send in two drives, it'd mean two of the SSDs get removed. I don't remember if the RAID controllers we have would support that mix but I remember it as being able to... or at least we have some that could do it.
I'd consider your drives as qualifying for the "pre-approved" you can get it back if you cancel territory but don't get mad if you contact us in 8 years and I've forgotten this policy essentially made up just for you at the moment. You'd have to still cover shipping/handling basically if you do want it back (shipping + pickup cost + packing + 15 minutes hands.)
Some notes on HW RAID controllers for Ryzen Dedis
This was a long section, I'll replace it with this: probably no working BBU. These are older controllers that used batteries, batteries expire. Even if we got brand new batteries there's no way to guarantee what shelf life we'd be getting on them, plus they're just generally not very good in the first place. If you really want it, pay us more for a proper controller and CacheVault.
Some notes on provisioning module for Dedicated servers and KVM dongles
I found a module we'll likely be using that supports our KVM switch to some degree, as well as likely supports our PDUs. So that means it might be more nicely integrated for KVM and "power cycling" if you want to call it that.
I also have a good update on this:
@VirMach said: I'm trying to modify the BIOS to have some important setting already set how we need it and trying to flash it with a programmer instead. This ends up saving a lot of time but at the same time it could end... not well. Basically the boards are super annoying with how they end up supporting the newer processors, I've already figured that part out, now we just need to replicate it quickly onto many boards. Otherwise I'm stuck putting a processor on, updating, removing, putting in new processor, and between the processor change the setting resets, and once it resets I lose VGA output. So then I have to connect an external GPU, and I don't have any GPUs that can just get enough power from PCIe right now, so I have to also bring in another power supply, reconnect everything, plug in the GPU, connect it through that, then change the setting, then save, then remove the GPU, rewire everything. Per server. And of course all the waiting for booting, the update to complete, any weird complications, etc.
After spending way too long on this, I now have a skill I did not think I'd ever need/use:
I have modified the BIOS chip and flashing them to our custom version using an external programmer. It hasn't saved any time so far and it's still not where I need it to be but several more hours and we might have our own custom Frankenstein BIOS. Okay no, we did get somewhere with them, worst case scenario I can avoid having to stick in a Ryzen 3000 series into every motherboard to flash it first but it's not good enough to avoid sticking in a GPU temporarily.
This is the part where I wrote out a lot of things but Chrome crashed and it didn't save it. I genuinely don't remember what it was about but several more sections and paragraphs.
@VirMach said: Some notes on HW RAID controllers for Ryzen Dedis
This was a long section, I'll replace it with this: probably no working BBU. These are older controllers that used batteries, batteries expire. Even if we got brand new batteries there's no way to guarantee what shelf life we'd be getting on them, plus they're just generally not very good in the first place. If you really want it, pay us more for a proper controller and CacheVault.
@VirMach said: Some notes on HW RAID controllers for Ryzen Dedis
This was a long section, I'll replace it with this: probably no working BBU. These are older controllers that used batteries, batteries expire. Even if we got brand new batteries there's no way to guarantee what shelf life we'd be getting on them, plus they're just generally not very good in the first place. If you really want it, pay us more for a proper controller and CacheVault.
🤦
Do JBOD + Soft RAID or configure it correctly to reduce risk, or pretend it doesn't exist. I went back and checked and it looks like I'm already being careful with how it's "advertised" as in I didn't mention RAID at all in the configuration listings even before posting the actual sale, and they also won't go in on our website either.
This isn't anything new, I bet I mentioned it on the OGF thread as well a few years ago.
Found it!
We did still add in RAID controllers without BBU and are plugging in the SSDs so people can get higher speeds on them without necessarily having to do RAID.
Wow 2021, what was that like a year ago?
EDIT -- Hmm, someone's selling bulk newer gen controllers pretty cheap, like the ones we paid $800-$1300 each for a few years ago. Very tempting but the last time I did that we didn't use them for 4 years I mean uhh, one year.
@taizi said: i think i will grab some storage offer,because they are raid10(really?),and can I buy 3 4TB plan and merge them together
Yes, RAID10. You can't merge them unless you just want to add 8TB to the 4TB without the extra memory, CPU, and IPv4. For the same price as buying three of course. No benefits to doing it other than big disk one place if needed. If there's demand I can see about adding another plan higher up or something.
I should really stop working on the BIOS for these Ryzen dedicated servers and go wrap everything else up.
@taizi said: also,any sneak peek about EPYC vps/dedi?
The dedicated server will likely be $59 to $69 a month, and 128GB to 256GB memory. With only a 1TB U.2 enterprise SSD (add $10-15 per month per 1TB additional that fits, or per 1TB extra on same drive.) I've probably said something similar or slightly different on it and it's not any more specific now.
For VPS, I don't want to overhype it but those really cheap high memory deals were planned on older generation Epyc for all of them except the obvious one (IPv6 only may go on dual Xeon or Gold.) The boards are finnicky on memory though as mentioned so they're not up to the higher memory amount yet, I did just buy about 10x 32GB today, buying more later. These sticks do cost like 33% (to realistically 66% more) though so it could end up affecting the plans for both dedicated server prices and/or memory amount and/or the VPS plans (with a bump up price, bump down memory, or bump down to dual E5's or Gold. We'll see. Well the gold's technically not a bump down but it feels like it, I can't see them as anything other than E5's pretending to be Gold.)
The early bird deals are really going to be early bird though, don't expect them to hang around forever or get restocked. I mean they probably could since it technically pays itself off and its portion of IPv4 and colocation, but I also want this location to pay off its really expensive network blend eventually.
I had more exact prices in my earlier draft, so if I did post that, then refer to that as it's more accurate: our network blend costs (MSRP) about $800 per 1Gbps for the amount we took in bulk. Compared to the cheapest option which is something like $70/1Gbps (MSRP.) $1200+ per 1Gbps not bulk. We of course did not pay MSRP for it, but I can also tell you it wasn't a 10x discount either. So we're really going for the premium network over single homing what we could have for much cheaper. That could have worked out too, and it would have meant Oklahoma would have been ultra cheap (cheaper than it is which is already pretty low) but I didn't want to turn it into our Buffalo. For example, it could have made the $25 a month Ryzen dedicated server early bird special be closer to $10 to 15 a month. Actually most of you would have probably preferred that but good thing I'm here to make questionable decisions be the grown up to make a potentially bad decision that helps keep it nice.
We'll probably go until at least some nodes are filled out and for some plans, there's not that many. Like a dozen or two of the higher end VPS plans could end up exhausting the early bird stock.
@VirMach said: If there's demand I can see about adding another plan higher up or something.
yes,higher plan better,and we can get more ram/cpu(compare to merge the disk only)
I'll see what we can do. It's just not set up that way but maybe it should be set up that way. Right now we have additional disk options for the 4TB plan because the order quantities on that are already low enough to where it doesn't make sense to have individual 8TB, 12TB, 16TB, and 32TB plans.
And going back to early birds, since it is RAID10, if a little over half a dozen people bought 12TB storage plans we'd be close to running out since we're reserving some space for our own stuff. For storage though I'm more likely to be able to extend the early bird prices if there's demand, just as pre-orders (with set date where we just refund everyone if it's not met.)
@VirMach said: The early bird deals are really going to be early bird though, don't expect them to hang around forever or get restocked
it makes me worried I can't grab one before others if I post benchmark result to MJJ forums
@VirMach said: I'll see what we can do. It's just not set up that way but maybe it should be set up that way. Right now we have additional disk options for the 4TB plan because the order quantities on that are already low enough to where it doesn't make sense to have individual 8TB, 12TB, 16TB, and 32TB plans.
what the specs of 4TB plan,if the cpu/memory/bandwidth is enough/good,I can just buy 4TB and choose additional disk options
who don't want more cpu/memory/bandwidth(even ips) for higher idle efficiency
@VirMach said: For the remainder I may just do 2600 or 3500 (with the 2600 you get full hyperthreading, with 3500 you don't but it performs the same. The 3500 should have better compatibility with the boards though.) These are the ones that will go in the $25 configuration so I'll let @Kris decide.
If that's the plan I go with though that might mean no 5500 version anymore for the $25 configuration and a slight delay on those until new CPUs come in, I might still be able to do a few of them with 2600's.
Whichever one allows me to purchase one quicker with a JBOD setup if possible. Like you said they both look similiar. I filled up Plex repo and playing with GetChannels now, and it needs a home, I'm riding my VM's at 80% disk and it's just what I need storage wise.
Comments
The control panel information has been updated!
VNC connection successful!
The gateway is configured correctly!
The gateway is globally accessible!
The main IP is set correctly!
Network configuration file is wrong, network interface should be ens3, auto-generated configuration is eth0, change it correctly!
Main IP is online!
Main IP is globally reachable!
wait what happened to virtfusion?
I bench YABS 24/7/365 unless it's a leap year.
This is a test from (almost) all my servers:
https://globalping.io/?measurement=zWFzj5EzGMeD2DTe
TL;DR. Seems quite good overall. Ping from home is roughly 126ms which is basically what you would expect (my ISP directly peers with Lumen so it pretty much goes to Paris and then directly to Oklahoma)
I've also noticed latency to Kansas City is a bit over the place (26ms/30ms). One of my boxes goes to Chicago and then directly to Oklahoma. Not that bad, but roughly 13ms or so increase.
My other box goes directly though Cogent until it hits Lumen basically. (Omaha - Denver - Tulsa)
EDIT: RIPE Atlas measurement (1000 US probes)
https://atlas.ripe.net/measurements/102150730/overview
MTR to Kuroit in Dallas, TX
MTR to ExtraVM in Dallas, TX
MTR to DartNode in Houston, TX
MTR to Kuroit in Ashburn, VA
MTR to LittleCreekHosting in Durham, NC
MTR to Crunchbits in Liberty Lake, WA
MTR to Crunchbits in Allentown, PA
MTR to GreenCloud in Chicago, IL
MTR to GreenCloud in Jacksonville, FL
MTR to HostHatch in Los Angeles, CA
MTR to HostCram in Salt Lake City, UT
I think VirtFusion was a "maybe" experiment
That one makes us look good, I like it. Time to figure out what's going on with Hivelocity Chicago (and T-Mobile Fiber Denver.)
Okay interesting, so Hivelocity Chicago is through Lumen, that's good. But I have to figure out how to make BGP to do that. I'm sure it'll be really easy. Wait, what's BGP? Ohhh... ohh... oh.
I haven't found a good route to Denver except that one, okay I've hacked the mainframe, I'm in. That one is Lumen as well. That's real good. Alright time to contact Lumen and see if they can make this easier.
You didn't say it enough.
It's not for these Ryzen servers. I mean I suppose it could be... I guess I never thought of it that way. I'd just need to trick WHMCS into using it for one location as I don't really want to make an entire OKC duplicate for each plan. For another product line on different servers, it made more sense. Or we could have sub-versions for Ryzen nodes.
Orrr.... I could finally get one of the other nodes up, it's crazy how backwards all our plans ended up going, these Ryzen were supposed to probably be last. Same with IPv6.
I'm sorry, what? I was not expecting that at all when opening that. Wow. How did you even... find that many cities in the world.
Okay I need to figure out a feasible way to use this data.
To KC 24.756 ms
1 _gateway (193.58.147.1) 2.695 ms 2.667 ms 2.648 ms
2 23.136.108.50 (23.136.108.50) 0.384 ms 0.364 ms 0.421 ms
3 1-1-c17-2.ear2.Tulsa2.Level3.net (4.43.131.85) 3.003 ms 2.982 ms 3.006 ms
4 * * *
5 * * 1299-3356-dal.sp.lumen.tech (4.68.74.166) 8.541 ms
6 * * dls-bb2-link.ip.twelve99.net (62.115.138.64) 8.520 ms
7 kanc-bb2-link.ip.twelve99.net (62.115.139.188) 13.507 ms 13.487 ms 13.311 ms
8 kanc-b3-link.ip.twelve99.net (62.115.113.81) 13.440 ms 13.572 ms 13.500 ms
9 nocix-ic-372038.ip.twelve99-cust.net (80.239.132.187) 13.077 ms 13.089 ms nocix-ic-372039.ip.twelve99-cust.net (62.115.147.181) 13.060 ms
10 100ge-0-38.edge-1.clay.mci.us.as32097.net (69.30.209.201) 13.417 ms 13.559 ms 100ge-103.dist-1.101.swift.mci.us.as32097.net (69.30.209.225) 24.634 ms
11 100ge-53-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.200) 12.991 ms 69.197.140.1 (69.197.140.1) 24.756 ms 24.771 ms
From KC 19.834 ms
1 _gateway (69.197.140.1) 0.167 ms 0.252 ms 0.244 ms
2 100ge-55-1.edge-2.101.swift.mci.us.as32097.net (69.30.209.224) 0.199 ms 0.213 ms 0.202 ms
3 kanc-b3-link.ip.twelve99.net (80.239.132.186) 0.598 ms 0.606 ms *
4 be5627.ccr22.mci01.atlas.cogentco.com (154.54.90.249) 1.370 ms 1.370 ms 1.611 ms
5 be5590.ccr31.oma02.atlas.cogentco.com (154.54.47.38) 4.576 ms * 4.843 ms
6 ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.317 ms be3802.ccr21.den01.atlas.cogentco.com (154.54.165.77) 15.064 ms ae5.5.ear2.Tulsa2.net.lumen.tech (4.69.227.145) 10.300 ms
7 be2766.ccr41.ord03.atlas.cogentco.com (154.54.46.178) 12.946 ms be2401.rcr21.b006467-6.den01.atlas.cogentco.com (154.54.88.42) 15.151 ms VIRTUAL-MAC.ear2.Tulsa2.Level3.net (4.43.131.86) 13.640 ms
8 orange.lax05.atlas.cogentco.com (154.54.10.2) 14.464 ms 14.429 ms 14.677 ms
9 193.58.147.1 (193.58.147.1) 19.861 ms 19.854 ms 19.834 ms
To Dallas 8.822 ms
1 _gateway (193.58.147.1) 4.330 ms 4.318 ms 4.314 ms
2 23.136.108.50 (23.136.108.50) 0.370 ms 0.394 ms 0.388 ms
3 1-1-c17-2.ear2.Tulsa2.Level3.net (4.43.131.85) 3.024 ms 3.016 ms 3.063 ms
4 ae1.3515.edge1.Dallas2.net.lumen.tech (4.69.209.114) 8.962 ms ae1.3509.edge2.Dallas2.net.lumen.tech (4.69.206.165) 8.244 ms 8.240 ms
5 * * *
6 be2763.ccr31.dfw01.atlas.cogentco.com (154.54.28.73) 8.877 ms 8.901 ms 8.941 ms
7 be8551.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.54.92.105) 9.382 ms be8553.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.54.95.29) 9.529 ms 9.632 ms
8 be4693.nr61.b059724-0.dfw01.atlas.cogentco.com (154.24.5.166) 9.591 ms 9.592 ms 9.465 ms
9 38.32.15.66 (38.32.15.66) 8.707 ms 15.827 ms 15.815 ms
10 64.44.158.1 (64.44.158.1) 8.882 ms 8.826 ms 8.822 ms
From Dallas 16.887 ms
1 1-148-44-64.reverse-dns (64.44.148.1) 0.172 ms 0.159 ms 0.176 ms
2 * * *
3 be4693.rcr71.b059724-0.dfw01.atlas.cogentco.com (154.24.5.165) 1.055 ms 1.170 ms 1.275 ms
4 be8553.ccr32.dfw01.atlas.cogentco.com (154.54.95.30) 1.249 ms 1.277 ms be8551.ccr31.dfw01.atlas.cogentco.com (154.54.92.106) 1.274 ms
5 be2763.ccr41.dfw03.atlas.cogentco.com (154.54.28.74) 1.242 ms be2764.ccr41.dfw03.atlas.cogentco.com (154.54.47.214) 1.144 ms 1.303 ms
6 lumen.dfw03.atlas.cogentco.com (154.54.10.18) 8.863 ms 1.620 ms 6.580 ms
7 ae2.13.ear2.Tulsa2.net.lumen.tech (4.69.136.210) 6.095 ms 6.112 ms 6.034 ms
8 VIRTUAL-MAC.ear2.Tulsa2.Level3.net (4.43.131.86) 8.885 ms 8.889 ms 8.900 ms
9 193.58.147.1 (193.58.147.1) 16.932 ms 16.910 ms 16.887 ms
That's interesting, and confusing. So is that the same routes I described except Cogent going from Omaha --> Denver --> Dallas --> OKC ends up being better than... oh I see it now, they handed it over to Telia from Dallas. And the Cogent one improved because of Telia in that case.
In my comparison, Telia wasn't involved I guess, just Cogent. It would've probably been helpful if I posted it, luckily I have the tab up. Oh nope it's not because of the Telia, it went to "Chicago" instead of "LAX" or.. is it the "Michigan" -- okay whatever I'm done playing Cogent placeholder hostname shuffle because it also geolocates to the that incorrect hostname still and I don't feel like triangulating each IP address.
I hope this isn't the standard way Cogent hands things off to Lumen in Dallas.
Here's some MTR data for the geographic locations I'm more interested in:
China(12)
#### Asia, Qingdao
DC514, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 11.247.239.22 | 0.0% | 2 | 0.9 | 0.8 | 0.7 | 0.9 | 0.1 |
| 2 | 11.73.5.129 | 0.0% | 2 | 2.2 | 3.1 | 2.2 | 4.0 | 1.3 |
| 3 | 10.54.143.237 | 0.0% | 2 | 6.3 | 3.5 | 0.7 | 6.3 | 4.0 |
| 4 | 116.251.117.222 | 0.0% | 2 | 1.1 | 1.1 | 1.1 | 1.2 | 0.0 |
| 5 | 116.251.113.45 | 0.0% | 2 | 0.8 | 0.9 | 0.8 | 0.9 | 0.0 |
| 6 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 7 | 140.249.132.73 | 0.0% | 2 | 2.0 | 2.0 | 2.0 | 2.0 | 0.0 |
| 8 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 9 | 202.97.99.161 | 50.0% | 2 | 13.9 | 13.9 | 13.9 | 13.9 | 0.0 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 12 | 202.97.92.74 | 0.0% | 2 | 147.7 | 147.7 | 147.7 | 147.7 | 0.0 |
| 13 | 4.15.125.53 | 0.0% | 2 | 146.9 | 146.9 | 146.8 | 146.9 | 0.1 |
| 14 | 4.69.227.145 | 0.0% | 2 | 185.4 | 186.9 | 185.4 | 188.5 | 2.2 |
| 15 | 4.43.131.86 | 0.0% | 2 | 191.6 | 191.5 | 191.3 | 191.6 | 0.2 |
| 16 | 193.58.147.1 | 0.0% | 2 | 239.4 | 223.1 | 206.8 | 239.4 | 23.1 |
HOST: qingdao-alibaba
#### Asia, Chengdu
DC523, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 11.202.9.2 | 50.0% | 2 | 0.6 | 0.6 | 0.6 | 0.6 | 0.0 |
| 2 | 11.73.2.205 | 0.0% | 2 | 0.9 | 1.0 | 0.9 | 1.0 | 0.0 |
| 3 | 11.88.170.249 | 0.0% | 2 | 0.7 | 0.7 | 0.6 | 0.7 | 0.0 |
| 4 | 45.112.221.42 | 0.0% | 2 | 1.4 | 1.4 | 1.3 | 1.4 | 0.1 |
| 5 | 116.251.119.138 | 0.0% | 2 | 4.1 | 3.0 | 2.0 | 4.1 | 1.5 |
| 6 | 101.207.253.225 | 0.0% | 2 | 1.6 | 1.6 | 1.6 | 1.6 | 0.0 |
| 7 | 119.6.197.225 | 0.0% | 2 | 8.4 | 5.7 | 3.1 | 8.4 | 3.7 |
| 8 | 219.158.106.249 | 0.0% | 2 | 39.4 | 39.3 | 39.2 | 39.4 | 0.1 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | 219.158.6.206 | 0.0% | 2 | 36.0 | 36.0 | 35.9 | 36.0 | 0.1 |
| 11 | 219.158.97.178 | 0.0% | 2 | 178.2 | 177.3 | 176.5 | 178.2 | 1.1 |
| 12 | 4.53.209.77 | 0.0% | 2 | 191.4 | 191.7 | 191.4 | 191.9 | 0.3 |
| 13 | 4.69.227.145 | 0.0% | 2 | 228.7 | 228.7 | 228.7 | 228.8 | 0.0 |
| 14 | 4.43.131.86 | 0.0% | 2 | 227.1 | 226.6 | 226.1 | 227.1 | 0.7 |
| 15 | 193.58.147.1 | 50.0% | 2 | 234.9 | 234.9 | 234.9 | 234.9 | 0.0 |
HOST: chengdu-alibaba
#### Asia, Guangzhou
DC522, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.231.251.10 | 0.0% | 2 | 0.5 | 0.6 | 0.5 | 0.7 | 0.1 |
| 2 | 11.73.14.177 | 0.0% | 2 | 0.4 | 0.4 | 0.4 | 0.4 | 0.0 |
| 3 | 11.73.25.173 | 0.0% | 2 | 0.4 | 0.4 | 0.4 | 0.4 | 0.0 |
| 4 | 11.73.48.129 | 0.0% | 2 | 0.7 | 0.7 | 0.7 | 0.7 | 0.0 |
| 5 | 11.94.57.209 | 0.0% | 2 | 0.8 | 0.9 | 0.8 | 0.9 | 0.1 |
| 6 | 10.54.220.161 | 0.0% | 2 | 0.8 | 0.8 | 0.8 | 0.8 | 0.0 |
| 7 | 11.94.155.73 | 0.0% | 2 | 1.2 | 1.2 | 1.2 | 1.2 | 0.0 |
| 8 | 157.148.24.41 | 0.0% | 2 | 1.2 | 1.2 | 1.2 | 1.2 | 0.0 |
| 9 | 120.80.137.89 | 0.0% | 2 | 3.1 | 2.7 | 2.4 | 3.1 | 0.6 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 12 | 219.158.30.26 | 0.0% | 2 | 30.1 | 29.5 | 28.9 | 30.1 | 0.8 |
| 13 | 219.158.8.182 | 0.0% | 2 | 33.5 | 32.1 | 30.8 | 33.5 | 1.9 |
| 14 | 219.158.116.238 | 0.0% | 2 | 181.8 | 181.8 | 181.7 | 181.8 | 0.0 |
| 15 | 4.53.209.77 | 0.0% | 2 | 161.9 | 160.9 | 160.0 | 161.9 | 1.3 |
| 16 | 4.69.227.145 | 0.0% | 2 | 196.0 | 195.9 | 195.9 | 196.0 | 0.1 |
| 17 | 4.43.131.86 | 0.0% | 2 | 201.0 | 202.1 | 201.0 | 203.3 | 1.6 |
| 18 | 193.58.147.1 | 0.0% | 2 | 207.3 | 203.9 | 200.6 | 207.3 | 4.7 |
HOST: guangzhou-alibaba
#### Asia, Hangzhou
DC208, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.12.208.214 | 0.0% | 2 | 1.8 | 1.8 | 1.8 | 1.8 | 0.0 |
| 2 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 3 | 10.255.101.109 | 0.0% | 2 | 1.6 | 2.6 | 1.6 | 3.7 | 1.4 |
| 4 | 11.94.128.130 | 0.0% | 2 | 3.1 | 8.5 | 3.1 | 13.9 | 7.7 |
| 5 | 10.102.42.37 | 0.0% | 2 | 3.5 | 3.6 | 3.5 | 3.6 | 0.0 |
| 6 | 115.236.101.74 | 0.0% | 2 | 3.8 | 3.8 | 3.8 | 3.8 | 0.0 |
| 7 | 220.191.200.57 | 0.0% | 2 | 4.1 | 4.1 | 4.1 | 4.1 | 0.0 |
| 8 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | 202.97.39.185 | 50.0% | 2 | 9.8 | 9.8 | 9.8 | 9.8 | 0.0 |
| 11 | 202.97.51.190 | 0.0% | 2 | 153.4 | 153.5 | 153.4 | 153.6 | 0.1 |
| 12 | 4.15.125.53 | 50.0% | 2 | 269.0 | 269.0 | 269.0 | 269.0 | 0.0 |
| 13 | 4.69.227.145 | 0.0% | 2 | 224.6 | 224.5 | 224.5 | 224.6 | 0.0 |
| 14 | 4.43.131.86 | 0.0% | 2 | 266.4 | 267.0 | 266.4 | 267.6 | 0.8 |
| 15 | 193.58.147.1 | 50.0% | 2 | 246.2 | 246.2 | 246.2 | 246.2 | 0.0 |
HOST: hangzhou-alibaba
#### Asia, Heyuan
DC521, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.4.144.118 | 0.0% | 2 | 0.8 | 0.9 | 0.8 | 0.9 | 0.0 |
| 2 | 10.4.144.1 | 0.0% | 2 | 1.4 | 1.3 | 1.2 | 1.4 | 0.1 |
| 3 | 10.54.231.153 | 0.0% | 2 | 0.9 | 0.8 | 0.7 | 0.9 | 0.1 |
| 4 | 10.54.230.158 | 0.0% | 2 | 1.2 | 1.2 | 1.2 | 1.2 | 0.0 |
| 5 | 157.119.193.54 | 0.0% | 2 | 6.8 | 6.8 | 6.7 | 6.8 | 0.0 |
| 6 | 11.94.143.130 | 0.0% | 2 | 9.1 | 8.8 | 8.6 | 9.1 | 0.3 |
| 7 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 8 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 9 | 119.147.220.129 | 0.0% | 2 | 11.6 | 11.6 | 11.6 | 11.6 | 0.0 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 12 | 202.97.27.238 | 0.0% | 2 | 161.9 | 191.5 | 161.9 | 221.2 | 41.9 |
| 13 | 4.15.125.53 | 0.0% | 2 | 165.1 | 164.7 | 164.3 | 165.1 | 0.6 |
| 14 | 4.69.227.145 | 0.0% | 2 | 201.6 | 200.0 | 198.4 | 201.6 | 2.2 |
| 15 | 4.43.131.86 | 0.0% | 2 | 208.3 | 217.6 | 208.3 | 226.9 | 13.2 |
| 16 | 193.58.147.1 | 0.0% | 2 | 211.1 | 209.3 | 207.5 | 211.1 | 2.6 |
HOST: heyuan-alibaba
#### Asia, Shenzhen
DC520, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.12.24.18 | 0.0% | 2 | 2.1 | 2.1 | 2.1 | 2.2 | 0.1 |
| 2 | 10.12.24.73 | 50.0% | 2 | 2.6 | 2.6 | 2.6 | 2.6 | 0.0 |
| 3 | 11.57.13.57 | 0.0% | 2 | 1.9 | 2.0 | 1.9 | 2.2 | 0.2 |
| 4 | 103.41.141.206 | 0.0% | 2 | 3.3 | 3.3 | 3.3 | 3.4 | 0.1 |
| 5 | 11.94.143.122 | 0.0% | 2 | 3.3 | 3.3 | 3.3 | 3.3 | 0.0 |
| 6 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 7 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 8 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | 202.97.22.122 | 0.0% | 2 | 422.5 | 292.1 | 161.8 | 422.5 | 184.3 |
| 12 | 4.15.125.53 | 0.0% | 2 | 161.6 | 165.4 | 161.6 | 169.2 | 5.4 |
| 13 | 4.69.227.145 | 0.0% | 2 | 205.3 | 205.1 | 204.9 | 205.3 | 0.3 |
| 14 | 4.43.131.86 | 0.0% | 2 | 204.9 | 205.5 | 204.9 | 206.1 | 0.9 |
| 15 | 193.58.147.1 | 0.0% | 2 | 208.4 | 242.1 | 208.4 | 275.7 | 47.6 |
HOST: shenzhen-alibaba
#### Asia, Shanghai
DC276, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 2 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 3 | 10.222.227.113 | 0.0% | 2 | 2.2 | 3.1 | 2.2 | 4.1 | 1.4 |
| 4 | 11.94.199.229 | 0.0% | 2 | 4.5 | 215.4 | 4.5 | 426.4 | 298.4 |
| 5 | 10.54.37.230 | 0.0% | 2 | 3.3 | 7.5 | 3.3 | 11.8 | 6.0 |
| 6 | 101.95.211.125 | 0.0% | 2 | 4.5 | 7.3 | 4.5 | 10.2 | 4.1 |
| 7 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 8 | 101.95.218.41 | 0.0% | 2 | 7.1 | 7.2 | 7.1 | 7.4 | 0.2 |
| 9 | 202.97.71.30 | 50.0% | 2 | 10.4 | 10.4 | 10.4 | 10.4 | 0.0 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | 202.97.52.250 | 50.0% | 2 | 146.9 | 146.9 | 146.9 | 146.9 | 0.0 |
| 12 | 4.15.125.53 | 0.0% | 2 | 213.6 | 213.8 | 213.6 | 214.1 | 0.3 |
| 13 | 4.69.227.145 | 0.0% | 2 | 226.4 | 226.4 | 226.4 | 226.4 | 0.0 |
| 14 | 4.43.131.86 | 0.0% | 2 | 219.0 | 218.9 | 218.7 | 219.0 | 0.2 |
| 15 | 193.58.147.1 | 0.0% | 2 | 259.4 | 259.0 | 258.5 | 259.4 | 0.6 |
HOST: shanghai-alibaba
#### Asia, Jining
DC517, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.123.208.6 | 0.0% | 2 | 1.3 | 1.4 | 1.3 | 1.5 | 0.2 |
| 2 | 10.123.208.73 | 0.0% | 2 | 2.8 | 6.2 | 2.8 | 9.6 | 4.8 |
| 3 | 10.54.223.253 | 0.0% | 2 | 1.4 | 1.5 | 1.4 | 1.7 | 0.2 |
| 4 | 117.49.33.214 | 0.0% | 2 | 1.4 | 1.3 | 1.2 | 1.4 | 0.1 |
| 5 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 6 | 116.251.112.201 | 0.0% | 2 | 8.5 | 8.5 | 8.5 | 8.5 | 0.0 |
| 7 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 8 | 36.110.245.209 | 0.0% | 2 | 10.0 | 10.0 | 10.0 | 10.0 | 0.0 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | 202.97.54.38 | 50.0% | 2 | 12.1 | 12.1 | 12.1 | 12.1 | 0.0 |
| 11 | 202.97.59.106 | 0.0% | 2 | 300.5 | 225.5 | 150.6 | 300.5 | 105.9 |
| 12 | 4.15.125.53 | 0.0% | 2 | 150.6 | 150.1 | 149.7 | 150.6 | 0.6 |
| 13 | 4.69.227.145 | 0.0% | 2 | 192.5 | 193.5 | 192.5 | 194.5 | 1.4 |
| 14 | 4.43.131.86 | 0.0% | 2 | 191.1 | 191.1 | 191.1 | 191.1 | 0.0 |
| 15 | 193.58.147.1 | 0.0% | 2 | 193.3 | 193.9 | 193.3 | 194.6 | 1.0 |
HOST: ulanqab-alibaba (Jining - Ulanqab is a common naming for datacenters in this region of Inner Mongolia, Jining being a city within Ulanqab league/prefecture)
#### Asia, Hohhot
DC329, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 11.55.35.62 | 0.0% | 2 | 0.8 | 0.7 | 0.6 | 0.8 | 0.1 |
| 2 | 11.73.1.193 | 0.0% | 2 | 3.6 | 4.3 | 3.6 | 5.0 | 1.0 |
| 3 | 11.141.77.58 | 0.0% | 2 | 0.5 | 0.5 | 0.5 | 0.5 | 0.0 |
| 4 | 116.251.105.146 | 0.0% | 2 | 0.6 | 0.6 | 0.5 | 0.6 | 0.0 |
| 5 | 103.52.73.22 | 0.0% | 2 | 15.7 | 15.7 | 15.7 | 15.7 | 0.0 |
| 6 | 116.251.112.197 | 0.0% | 2 | 12.8 | 12.7 | 12.6 | 12.8 | 0.2 |
| 7 | 106.38.196.85 | 0.0% | 2 | 13.3 | 13.2 | 13.1 | 13.3 | 0.2 |
| 8 | 36.110.244.1 | 0.0% | 2 | 14.4 | 16.5 | 14.4 | 18.7 | 3.1 |
| 9 | 202.97.94.186 | 0.0% | 2 | 22.5 | 20.3 | 18.1 | 22.5 | 3.1 |
| 10 | 202.97.12.134 | 0.0% | 2 | 15.8 | 16.0 | 15.8 | 16.1 | 0.2 |
| 11 | 202.97.59.106 | 0.0% | 2 | 155.2 | 156.1 | 155.2 | 157.1 | 1.3 |
| 12 | 4.15.125.53 | 0.0% | 2 | 155.8 | 157.9 | 155.8 | 160.1 | 3.1 |
| 13 | 4.69.227.145 | 0.0% | 2 | 192.4 | 192.4 | 192.4 | 192.5 | 0.1 |
| 14 | 4.43.131.86 | 0.0% | 2 | 197.0 | 197.3 | 197.0 | 197.6 | 0.4 |
| 15 | 193.58.147.1 | 0.0% | 2 | 196.0 | 196.2 | 196.0 | 196.5 | 0.3 |
HOST: huhehaote-alibaba
#### Asia, Beijing
DC207, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.36.68.38 | 50.0% | 2 | 2.2 | 2.2 | 2.2 | 2.2 | 0.0 |
| 2 | 11.73.8.170 | 0.0% | 2 | 2.3 | 2.3 | 2.3 | 2.4 | 0.1 |
| 3 | 10.102.252.66 | 0.0% | 2 | 2.3 | 2.3 | 2.2 | 2.3 | 0.0 |
| 4 | 10.102.251.190 | 50.0% | 2 | 4.3 | 4.3 | 4.3 | 4.3 | 0.0 |
| 5 | 45.112.216.138 | 0.0% | 2 | 4.6 | 4.7 | 4.6 | 4.7 | 0.1 |
| 6 | 36.110.217.9 | 50.0% | 2 | 4.9 | 4.9 | 4.9 | 4.9 | 0.0 |
| 7 | 36.110.245.189 | 0.0% | 2 | 5.8 | 5.9 | 5.8 | 6.0 | 0.2 |
| 8 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | 202.97.51.193 | 0.0% | 2 | 145.4 | 145.4 | 145.3 | 145.4 | 0.1 |
| 11 | 4.15.125.53 | 0.0% | 2 | 144.2 | 151.9 | 144.2 | 159.6 | 10.9 |
| 12 | 4.69.227.145 | 0.0% | 2 | 182.6 | 182.7 | 182.6 | 182.8 | 0.1 |
| 13 | 4.43.131.86 | 0.0% | 2 | 187.2 | 187.3 | 187.2 | 187.3 | 0.1 |
| 14 | 193.58.147.1 | 0.0% | 2 | 189.4 | 204.8 | 189.4 | 220.2 | 21.8 |
HOST: beijing-alibaba.perfops.org
#### Asia, Nanjing
DC518, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 2 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 3 | 11.73.1.65 | 50.0% | 2 | 0.7 | 0.7 | 0.7 | 0.7 | 0.0 |
| 4 | 11.94.208.225 | 0.0% | 2 | 0.7 | 0.5 | 0.4 | 0.7 | 0.2 |
| 5 | 10.54.132.102 | 0.0% | 2 | 9.7 | 9.6 | 9.6 | 9.7 | 0.1 |
| 6 | 116.251.113.217 | 0.0% | 2 | 8.3 | 18.6 | 8.3 | 28.9 | 14.6 |
| 7 | 180.163.38.81 | 0.0% | 2 | 9.6 | 9.7 | 9.6 | 9.8 | 0.1 |
| 8 | 101.95.209.69 | 50.0% | 2 | 20.6 | 20.6 | 20.6 | 20.6 | 0.0 |
| 9 | 101.95.218.73 | 0.0% | 2 | 10.5 | 10.6 | 10.5 | 10.7 | 0.1 |
| 10 | 202.97.94.237 | 0.0% | 2 | 24.9 | 25.5 | 24.9 | 26.1 | 0.8 |
| 11 | 202.97.12.186 | 0.0% | 2 | 13.6 | 13.7 | 13.6 | 13.8 | 0.1 |
| 12 | 202.97.51.190 | 0.0% | 2 | 184.7 | 184.7 | 184.7 | 184.7 | 0.0 |
| 13 | 4.15.125.53 | 50.0% | 2 | 183.1 | 183.1 | 183.1 | 183.1 | 0.0 |
| 14 | 4.69.227.145 | 0.0% | 2 | 222.6 | 222.5 | 222.4 | 222.6 | 0.1 |
| 15 | 4.43.131.86 | 50.0% | 2 | 230.6 | 230.6 | 230.6 | 230.6 | 0.0 |
| 16 | 193.58.147.1 | 0.0% | 2 | 230.2 | 230.4 | 230.2 | 230.7 | 0.3 |
HOST: nanjing-alibaba
#### Asia, Zhangjiakou
DC516, AS37963
| Hop | HOST | Loss% | Snt | Last | Avg | Best | Wrst | StDev |
|:----|:-----------------|------:|----:|------:|------:|------:|------:|------:|
| 1 | 10.130.116.42 | 0.0% | 2 | 1.1 | 1.1 | 1.1 | 1.2 | 0.1 |
| 2 | 11.73.30.66 | 50.0% | 2 | 0.9 | 0.9 | 0.9 | 0.9 | 0.0 |
| 3 | 10.216.175.206 | 0.0% | 2 | 1.5 | 1.6 | 1.5 | 1.8 | 0.2 |
| 4 | 10.255.161.134 | 0.0% | 2 | 1.2 | 1.3 | 1.2 | 1.4 | 0.1 |
| 5 | 10.54.111.174 | 0.0% | 2 | 5.7 | 5.9 | 5.7 | 6.0 | 0.2 |
| 6 | 10.102.155.34 | 0.0% | 2 | 5.7 | 5.7 | 5.7 | 5.7 | 0.0 |
| 7 | 106.38.196.177 | 0.0% | 2 | 7.5 | 7.5 | 7.5 | 7.6 | 0.1 |
| 8 | 36.110.245.201 | 0.0% | 2 | 7.6 | 7.6 | 7.6 | 7.7 | 0.1 |
| 9 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 10 | ??? | 100.0 | 2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 11 | 202.97.51.193 | 0.0% | 2 | 146.7 | 158.9 | 146.7 | 171.1 | 17.2 |
| 12 | 4.15.125.53 | 0.0% | 2 | 153.3 | 149.8 | 146.2 | 153.3 | 5.0 |
| 13 | 4.69.227.145 | 0.0% | 2 | 187.1 | 187.3 | 187.1 | 187.6 | 0.4 |
| 14 | 4.43.131.86 | 0.0% | 2 | 190.5 | 190.0 | 189.5 | 190.5 | 0.7 |
| 15 | 193.58.147.1 | 0.0% | 2 | 194.4 | 192.0 | 189.5 | 194.4 | 3.4 |
HOST: zhangjiakou-alibaba
Japan(13)
#### Asia, Tokyo
DC409, AS36236
HOST: tokyo-netactuate.perfops.org
#### Asia, Osaka
DC368, AS396982
HOST: osaka-google
#### Asia, Tokyo
DC441, AS16509
HOST: tokyo-2-aws
#### Asia, Tokyo
DC315, AS25108
HOST: tokyo-dediserve.perfops.org
#### Asia, Osaka
DC498, AS20473
HOST: osaka-vultr
#### Asia, Tokyo
DC545, AS63949
HOST: jptyo02-mon02
#### Asia, Tokyo
DC565, AS36351
HOST: jptyo01-mon02
#### Asia, Tokyo
DC564, AS36351
HOST: jptyo01-mon01
#### Asia, Tokyo
DC6, AS20473
HOST: tokyo-vultr.perfops.org
#### Asia, Tokyo
DC367, AS396982
HOST: tokyo-google
#### Asia, Tokyo
DC544, AS63949
HOST: jptyo02-mon01
#### Asia, Tokyo
DC118, AS63949
HOST: tokyo-linode.perfops.org
#### Asia, Osaka
DC142, AS134835
HOST: osaka-starrydns
Singapore(10)
#### Singapore
DC440, AS16509
HOST: singapore-2-aws.perfops.org
#### Singapore
DC331, AS202053
HOST: singapore-upcloud.perfops.org
#### Singapore
DC411, AS36236
HOST: singapore-netactuate
#### Singapore
DC60, AS38001
HOST: singapore-dediserve.perfops.org
#### Singapore
DC370, AS396982
HOST: singapore-google
#### Singapore
DC29, AS7489
HOST: singapore-hostus
#### Singapore
DC546, AS63949
HOST: sgsin02-mon01
#### Singapore
DC332, AS14061
HOST: singapore-digitalocean
#### Singapore
DC547, AS63949
HOST: sgsin02-mon02
#### Singapore
DC163, AS20473
HOST: singapore-vultr.perfops.org
to guangzhou - china mobile
I'm uncertain if my ISP limited single-parallel speed to 8Mbps, as iperf reaches 50Mbps when using 8 parallels.
ignore the iperf, speedtest with single-thread is excellent
Europe(48)
Europe, Bratislava
DC104, AS201924
HOST: bratislava-enahost.perfops.org
Europe, Frankfurt am Main
DC137, AS9009
HOST: frankfurt-edis.perfops.org
Europe, Amsterdam
DC344, AS13213
HOST: amsterdam-vpsnet.perfops.org
Europe, Limassol
DC229, AS197648
HOST: limassol-onapp.perfops.org
Europe, Turin
DC528, AS396982
(empty response)
Europe, Vienna
DC538, AS57169
HOST: atvie01-mon02
Europe, Varna
DC171, AS56630
HOST: varna-edis.perfops.org
Europe, Luxembourg
DC295, AS202422
HOST: luxembourg-gcore.perfops.org
Europe, Paris
DC418, AS36236
HOST: paris-netactuate
Europe, Vienna
DC338, AS61098
HOST: vienna-exoscale
Europe, London
DC543, AS63949
HOST: gblon01-mon02
Europe, Amsterdam
DC25, AS43350
HOST: amsterdam-edis.perfops.org
Europe, Copenhagen
DC570, AS136258
HOST: dkcph01-mon01
Europe, Paris
DC5, AS20473
HOST: paris-vultr.perfops.org
Europe, Vienna
DC537, AS57169
HOST: atvie01-mon01
Europe, Zurich
DC300, AS61098
HOST: zurich-exoscale
Europe, Geneve
DC174, AS61098
HOST: geneva-exoscale
Europe, London
DC423, AS36236
HOST: london-netactuate.perfops.org
Europe, Belgrade
DC475, AS136258
HOST: belgrade-oneprovider.perfops.org
Europe, Budapest
DC180, AS62214
HOST: vm3302
Europe, Florence
DC474, AS49367
HOST: florence-oneprovider.perfops.org
Europe, Helsinki
DC477, AS51765
HOST: helsinki-oneprovider.perfops.org
Europe, London
DC19, AS59764
HOST: london-atlantic.perfops.org
Europe, Frankfurt am Main
DC374, AS396982
HOST: frankfurt-google
Europe, Brussels
DC476, AS9009
HOST: brussels-oneprovider.perfops.org
Europe, Riga
DC89, AS52173
HOST: riga-yourserver
Europe, Amsterdam
DC153, AS50673
HOST: amsterdam-dediserve
Europe, Dronten
DC252, AS208258
HOST: dronten-crowncloud
Europe, London
DC34, AS62240
HOST: london-hostus
Europe, Chisinau
DC200, AS39798
HOST: chisinau-mivocloud.perfops.org
Europe, London
DC4, AS20473
HOST: london-vultr.perfops.org
Europe, Amsterdam
DC532, AS14061
HOST: nlams02-testmon01
Europe, Frankfurt am Main
DC155, AS44066
HOST: frankfurt-dediserve.perfops.org
Europe, Marseille
DC525, AS138915
HOST: marseille-lightnode.perfops.org
Europe, Vilnius
DC76, AS212531
HOST: vilnius-time4vps.perfops.org
Europe, Sandefjord
DC169, AS56655
HOST: sandefjord-terrahost.perfops.org
Europe, Madrid
DC421, AS36236
HOST: madrid-netactuate
Europe, Amsterdam
DC533, AS14061
HOST: nlams02-testmon02
Europe, London
DC542, AS63949
HOST: gblon01-mon01
Europe, Zurich
DC376, AS396982
HOST: zurich-google
Europe, Frankfurt am Main
DC426, AS36236
HOST: frankfurt-netactuate
Europe, Madrid
DC45, AS136258
HOST: madrid-oneprovider.perfops.org
Europe, Manchester
DC500, AS20473
HOST: manchester-vultr
Europe, Vienna
DC26, AS57169
HOST: vienna-edis
Europe, Johvi
DC100, AS198068
HOST: johvi-fastvps.perfops.org
Europe, Stockholm
DC568, AS136258
HOST: sesto01-mon01
Europe, Vienna
DC154, AS203833
HOST: vienna-dediserve
Europe, Frankfurt am Main
DC461, AS61098
HOST: frankfurt-exoscale
Europe, Stockholm
DC569, AS136258
HOST: sesto01-mon02
I see you guys have been working. Nice.
DC Lat / Lng? So i can add this to NMS map :-)
From NE Scotland:
3 ae50-ner001.msp.as13285.net (78.144.1.33) 13.928 ms 14.179 ms 14.436 ms
4 ae50-scr001-msp.as13285.net (78.144.1.32) 14.737 ms 15.337 ms 15.668 ms
5 ae60-scr101.thw.as13285.net (78.144.1.110) 21.508 ms 22.113 ms 29.658 ms
6 ae54.edge7.london1.level3.net (217.163.47.149) 30.129 ms 23.205 ms 28.596 ms
7 ae5.5.ear2.tulsa2.net.lumen.tech (4.69.227.145) 118.941 ms 119.222 ms 119.411 ms
8 virtual-mac.ear2.tulsa2.level3.net (4.43.131.86) 122.410 ms 122.703 ms 122.947 ms
9 23.136.108.51 (23.136.108.51) 123.682 ms 123.938 ms 131.552 ms
10 193.58.147.xxx (193.58.147.xxx) 123.228 ms 120.955 ms 121.567 ms
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural.
It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
Some high quality Lumen (albeit loves to backhaul through DFW) from all tests I've done.
Do we need to play some sort of Saw type game to get that $25 dedi @VirMach ? This would be the perfect 'not just another DC in the tri-state area' for backups and I need one more bare metal server if it's possible. Let me know who / what I need to sacrifice to get in on that deal.
Ring the 308 buzzer every day until you get $25 dedi with 1200GB NVMe.
We accept Karma donations for the last flan. 🍮 affbrr
I want my Virmach servers in Dallas please.
Now this is confusing (for me): my ATL hostname has a MIA node name, yet is now in OKC.
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural.
It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
@Virnach, I like it so far, can I ask one more migration by opening same form of ticket?
Yes
No .
I asked one more anyway, Ticket #115965
Thanks @VirMach for adding the extra bandwidth.
But I won't be using that bandwidth for a while.
So I set up an unofficial OKC speed test service
For those who don't have an OKC server yet and are interested in that location try it:http://193.58.147.138/test/speedtest/
For privacy reasons, I will not record the IPs of visitors (will uniformly use 0.0.0.0 instead), but will record download speeds, upload speeds, and ISP sources (not very accurate).
All data will be deleted at the end of the next month or before the service stops, and this service will stop at the end of the next month or when the bandwidth runs out.
Hope you guys like it.
(Please don't kill my server
)
I bench YABS 24/7/365 unless it's a leap year.
@VirMach
From WHMCS, VNC works as it should. Change host name works.
From SolusVM, reconfigure networking works on template installed VPS.
I did not found any bugs in the control panels so far.
Will try more when reinstall is available.
Network is nice and steady.
question:why everytime start i need to wait 1min30s
answer: check logs.
However I recall this feature only in Ubuntu waiting for network and you needed to tweak some settings (netplan (c)), but yours is by disk - checking something?
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
Check your fstab. It might be looking for something by UUID that doesn't exist.
answer: the grease needs to warm up to allow the rust to spin on the disc.
In stasis until the shitposting stops/abates.
Than=compare;then=sequence:brought=bring;bought=buy:staffs=pile of sticks:informations/infos=no plural.
It wisnae me! A big boy done it and ran away. || NVMe2G for life! until death (the end is nigh).
All accurate in a 630 mile radius
Order them and I can unplug it and drive it down to Dallas for you, in a self-storage unit.
I've spoken with our "solution engineer" at Lumen.
So Lumen's closest POP is Tulsa even though they have lines going up out of Oklahoma City. I assume in some specific cases this could mean theoretically, one route that looks closer might actually be further away since it has to come in from Tulsa but we've already confirmed for Denver and Chicago that there should be no reason for it to need to backhaul to Dallas.
They're going to speak with a "WAN Architect" to come up with some suggested changes.
35.4668473, -97.6492831
I'm trying to modify the BIOS to have some important setting already set how we need it and trying to flash it with a programmer instead. This ends up saving a lot of time but at the same time it could end... not well. Basically the boards are super annoying with how they end up supporting the newer processors, I've already figured that part out, now we just need to replicate it quickly onto many boards. Otherwise I'm stuck putting a processor on, updating, removing, putting in new processor, and between the processor change the setting resets, and once it resets I lose VGA output. So then I have to connect an external GPU, and I don't have any GPUs that can just get enough power from PCIe right now, so I have to also bring in another power supply, reconnect everything, plug in the GPU, connect it through that, then change the setting, then save, then remove the GPU, rewire everything. Per server. And of course all the waiting for booting, the update to complete, any weird complications, etc.
I should've just done 3000 series to be honest. All of this is because someone at ASRock did the timings incorrectly for something for this specific AMD processor and just "assumed" all 5000 series are the same. I kind of agree with them though, AMD shouldn't have had sub-series or whatever you want to call it for processors, just call it something else, don't pretend it's in the same line. Even worse, within the same series they specifically made for being APUs, they went back and disabled the integrated GPU.
For the remainder I may just do 2600 or 3500 (with the 2600 you get full hyperthreading, with 3500 you don't but it performs the same. The 3500 should have better compatibility with the boards though.) These are the ones that will go in the $25 configuration so I'll let @Kris decide.
If that's the plan I go with though that might mean no 5500 version anymore for the $25 configuration and a slight delay on those until new CPUs come in, I might still be able to do a few of them with 2600's.
Default SolusVM template. I had some idea for fixing it in our own template version but those never actually get finished.
OKC Update
Yesterday I got additional servers fully ready. The KVM switch is also in place. Rails also came in for some servers. The dongles also came in I think for KVM switch. Cox Communications also asked us for the 3rd time if we want single mode or multi-mode fiber so I've answered them again.
I'm probably going to get some other servers ready, and go in to the DC tonight with them and then maybe redo the cable management and organize things a bit now that it's taking shape. This might mean the first unscheduled maintenance so beta testers may go down for a bit. I also need to change the DAC, I had to use one that isn't ideal length. Except I don't even think I have the new one yet so that might be our second unscheduled maintenance.
IP addresses may also change or I might migrate everyone to a different node.
The storage node has an SSD now, but this board also has some weird VGA issue, I might need to update the BIOS or just fix IPMI do it through that, it needs operating system installed.
Epyc nodes, we have two more that are "ready-ish" and another two coming in, with a sixth one being built-ish. These all have finnicky motherboards that won't accept the memory I have, so I'm ordering additional memory for them as well.
One big issue at the facility is that (a) the WiFi doesn't work which isn't uncommon, and (b) my cell reception has gotten to the level of being completely unusable. So I'll try to remember to take down a WiFi router.
Then I'll probably spend the rest of the time today or tomorrow setting up VirtFusion and/or dedicated server provisioning module of sorts, maybe also setting up the controls for the PDUs. But I'd probably say most my time is going to be spent staring at the cabinets and wondering what should go where, like for way too long instead of actually doing it, so inevitability lots of shuffling positions.
The rest of the weekend, likely no sale, but spring cleaning. I already have too many boxes and too many things I took out of shelves, I need to re-organize inventory. I'm still in OKC, my flight got cancelled so for now the trip I was going to take has been cancelled. So if anything goes wrong then I should be able to fix it pretty quickly after my monitoring system alerts me (you guys on this LES thread at the moment.)
I'll try to process any additional requests in daily batches. Yes, you guys can keep putting them in.
Wait wait, we're doing dedis again? Oh man, I picked the perfect time to check in again, I definitely want to get in on this. My old dedi was one of those insanely unsustainable deals, I think 225/yr for 32gb/240gb, and you went so far above and beyond already with everything that happened through your dedi ordeals. I'd love to snag one of these, especially with the option of sending in my own drives. I've got a pair of 14tb wd reds I bought for a NAS build that never happened, and rather that keep them sitting around gathering dust, this seems like the perfect use for them.
Yep, we can do that. Just put in a ticket. So your equivalent based on what you said would be:
And it'd be $225 a year or whatever you mentioned instead of the $35 a month. If you want to send in two drives, it'd mean two of the SSDs get removed. I don't remember if the RAID controllers we have would support that mix but I remember it as being able to... or at least we have some that could do it.
I'd consider your drives as qualifying for the "pre-approved" you can get it back if you cancel territory but don't get mad if you contact us in 8 years and I've forgotten this policy essentially made up just for you at the moment. You'd have to still cover shipping/handling basically if you do want it back (shipping + pickup cost + packing + 15 minutes hands.)
Some notes on HW RAID controllers for Ryzen Dedis
This was a long section, I'll replace it with this: probably no working BBU. These are older controllers that used batteries, batteries expire. Even if we got brand new batteries there's no way to guarantee what shelf life we'd be getting on them, plus they're just generally not very good in the first place. If you really want it, pay us more for a proper controller and CacheVault.
Some notes on provisioning module for Dedicated servers and KVM dongles
I found a module we'll likely be using that supports our KVM switch to some degree, as well as likely supports our PDUs. So that means it might be more nicely integrated for KVM and "power cycling" if you want to call it that.
I also have a good update on this:
After spending way too long on this, I now have a skill I did not think I'd ever need/use:
I have modified the BIOS chip and flashing them to our custom version using an external programmer. It hasn't saved any time so far and it's still not where I need it to be but several more hours and we might have our own custom Frankenstein BIOS. Okay no, we did get somewhere with them, worst case scenario I can avoid having to stick in a Ryzen 3000 series into every motherboard to flash it first but it's not good enough to avoid sticking in a GPU temporarily.
This is the part where I wrote out a lot of things but Chrome crashed and it didn't save it. I genuinely don't remember what it was about but several more sections and paragraphs.
🤦
Do JBOD + Soft RAID or configure it correctly to reduce risk, or pretend it doesn't exist. I went back and checked and it looks like I'm already being careful with how it's "advertised" as in I didn't mention RAID at all in the configuration listings even before posting the actual sale, and they also won't go in on our website either.
This isn't anything new, I bet I mentioned it on the OGF thread as well a few years ago.
Found it!
Wow 2021, what was that like a year ago?
EDIT -- Hmm, someone's selling bulk newer gen controllers pretty cheap, like the ones we paid $800-$1300 each for a few years ago. Very tempting but the last time I did that we didn't use them for 4 years I mean uhh, one year.
i think i will grab some storage offer,because they are raid10(really?),and can I buy 3 4TB plan and merge them together
also,any sneak peek about EPYC vps/dedi?
Yes, RAID10. You can't merge them unless you just want to add 8TB to the 4TB without the extra memory, CPU, and IPv4. For the same price as buying three of course. No benefits to doing it other than big disk one place if needed. If there's demand I can see about adding another plan higher up or something.
I should really stop working on the BIOS for these Ryzen dedicated servers and go wrap everything else up.
The dedicated server will likely be $59 to $69 a month, and 128GB to 256GB memory. With only a 1TB U.2 enterprise SSD (add $10-15 per month per 1TB additional that fits, or per 1TB extra on same drive.) I've probably said something similar or slightly different on it and it's not any more specific now.
For VPS, I don't want to overhype it but those really cheap high memory deals were planned on older generation Epyc for all of them except the obvious one (IPv6 only may go on dual Xeon or Gold.) The boards are finnicky on memory though as mentioned so they're not up to the higher memory amount yet, I did just buy about 10x 32GB today, buying more later. These sticks do cost like 33% (to realistically 66% more) though so it could end up affecting the plans for both dedicated server prices and/or memory amount and/or the VPS plans (with a bump up price, bump down memory, or bump down to dual E5's or Gold. We'll see. Well the gold's technically not a bump down but it feels like it, I can't see them as anything other than E5's pretending to be Gold.)
yes,higher plan better,and we can get more ram/cpu(compare to merge the disk only)
The early bird deals are really going to be early bird though, don't expect them to hang around forever or get restocked. I mean they probably could since it technically pays itself off and its portion of IPv4 and colocation, but I also want this location to pay off its really expensive network blend eventually.
I had more exact prices in my earlier draft, so if I did post that, then refer to that as it's more accurate: our network blend costs (MSRP) about $800 per 1Gbps for the amount we took in bulk. Compared to the cheapest option which is something like $70/1Gbps (MSRP.) $1200+ per 1Gbps not bulk. We of course did not pay MSRP for it, but I can also tell you it wasn't a 10x discount either. So we're really going for the premium network over single homing what we could have for much cheaper. That could have worked out too, and it would have meant Oklahoma would have been ultra cheap (cheaper than it is which is already pretty low) but I didn't want to turn it into our Buffalo. For example, it could have made the $25 a month Ryzen dedicated server early bird special be closer to $10 to 15 a month. Actually most of you would have probably preferred that but good thing I'm here to make questionable decisions be the grown up to make a potentially bad decision that helps keep it nice.
We'll probably go until at least some nodes are filled out and for some plans, there's not that many. Like a dozen or two of the higher end VPS plans could end up exhausting the early bird stock.
I'll see what we can do. It's just not set up that way but maybe it should be set up that way. Right now we have additional disk options for the 4TB plan because the order quantities on that are already low enough to where it doesn't make sense to have individual 8TB, 12TB, 16TB, and 32TB plans.
And going back to early birds, since it is RAID10, if a little over half a dozen people bought 12TB storage plans we'd be close to running out since we're reserving some space for our own stuff. For storage though I'm more likely to be able to extend the early bird prices if there's demand, just as pre-orders (with set date where we just refund everyone if it's not met.)
it makes me worried I can't grab one before others if I post benchmark result to MJJ forums
what the specs of 4TB plan,if the cpu/memory/bandwidth is enough/good,I can just buy 4TB and choose additional disk options
who don't want more cpu/memory/bandwidth(even ips) for higher idle efficiency
Virtfuuuuuuuuuuuuuuuuuuuuuuusion
I bench YABS 24/7/365 unless it's a leap year.
Whichever one allows me to purchase one quicker with a JBOD setup if possible. Like you said they both look similiar. I filled up Plex repo and playing with GetChannels now, and it needs a home, I'm riding my VM's at 80% disk and it's just what I need storage wise.
I must have missed something - is there an order link to early bird storage ?
Get your FREE VPS if you develop Open Source software
I want a dedi. I don't need it. But I want it. As usual.
Me too...
i dont want a dedi. i want 100 vps
I bench YABS 24/7/365 unless it's a leap year.