Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


★ VirMach ★ RYZEN ★ NVMe ★★ $8.88/YR- 384MB ★★ $21.85/YR- 2.5GB ★ Instant ★ Japan Pre-order ★ & More - Page 199
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

★ VirMach ★ RYZEN ★ NVMe ★★ $8.88/YR- 384MB ★★ $21.85/YR- 2.5GB ★ Instant ★ Japan Pre-order ★ & More

1196197199201202339

Comments

  • AlwaysSkintAlwaysSkint Member
    edited May 2022

    @RecD said: So many bugs these days.

    Nothing new in this. A consequence of too many changes in hardware/firmware/software within a short timeframe. The primary reason for large corporations sticking with often deemed "obsolete/deprecated" (semi-)proven technology.
    (Who said, IBM system 36/38, out there? I can't recall the model of NCR we had, with "Genisis" patched on top of it, for legacy Cobol.)

    Thanked by 2FrankZ ZA_capetown
  • FrankZFrankZ Veteran
    edited May 2022

    @AlwaysSkint said: Oh, oh, here comes the Dremel

    Funny that. You though Dremel. I though band saw.

    I was talking about writing in RPG III for system 38/ AS400 the other day.
    I remember the days when everything I wrote was designed to run for 30 years.
    Marriott hotel's system ran on an AS400 as late as the early 2000's.
    Things change so fast now that I sometimes wonder how they ever get it to be stable.

  • VirMachVirMach Member, Patron Provider

    @tonyairicon said:

    @VirMach said:
    Fixed a bug with anti-abuse script which was causing some of the overloading/abuse problems by not catching them correctly on some Ryzen nodes.

    I did the migration and then the bandwidth seems to be having issues, 15000MB became 15000TB! As much as I would love to have that much bandwidth, it's probably really a bug.

    Description My migration payment ticket is Ticket #548489, you can check the bug according to this

    Bandwidth 0 B of 15000 TB Used / 15000 TB Free

    Looks like either we give 1GB bandwidth or 15,000TB. Nothing in between! We'll have it corrected but we're not too worried about it going this way, it's not like anyone is going to use 15,000TB. Don't prove me wrong.

  • MikePTMikePT Moderator, Patron Provider, Veteran

    @VirMach said:

    @tonyairicon said:

    @VirMach said:
    Fixed a bug with anti-abuse script which was causing some of the overloading/abuse problems by not catching them correctly on some Ryzen nodes.

    I did the migration and then the bandwidth seems to be having issues, 15000MB became 15000TB! As much as I would love to have that much bandwidth, it's probably really a bug.

    Description My migration payment ticket is Ticket #548489, you can check the bug according to this

    Bandwidth 0 B of 15000 TB Used / 15000 TB Free

    Looks like either we give 1GB bandwidth or 15,000TB. Nothing in between! We'll have it corrected but we're not too worried about it going this way, it's not like anyone is going to use 15,000TB. Don't prove me wrong.

    That'd be interesting.

  • AlwaysSkintAlwaysSkint Member
    edited May 2022

    @FrankZ said: Things change so fast now that I sometimes wonder how they ever get it to be stable.

    Precisely.
    (I had the dubious distinction of being 1st to setup AS/400 PC Support software suite, used in a commercial environment, in Scotland, perhaps UK. Token Ring, network shares and all that palaver. It was the desk jockeys that pushed out the COBOL/RPG III. At that time UNIX was just a glimmer in my eye - ooh, that looks interesting. Re: Marriott=y2k issues? IIRC, AIX was coded to go boom in 2018, back in '99.)

    Thanked by 2FrankZ ZA_capetown
  • @FrankZ said: You thought Dremel. I thought band saw..

    Heh. :p

  • VirMachVirMach Member, Patron Provider

    @FrankZ said: Things change so fast now that I sometimes wonder how they ever get it to be stable.

    Thanked by 2FrankZ TimboJones
  • VirMachVirMach Member, Patron Provider

    @LiliLabs just tagging you because we discovered something today relating to the X470 motherboards, so I thought you'd want to hear it. I said earlier this week that it was strange how all of them would test well, and then at the datacenter have problems to where DC hands couldn't even get IPMI set up, frozen at the screen.

    It turns out those are all X470 because of some bug in the BIOS. ASRock said not to update to the latest BIOS for 3000 series so I tried to listen... well, the old version apparently has this weird issue where on reboot when trying to get into BIOS it just freezes. If IPMI is set up, you can go in and use it just fine, but the problem is if they need to get into the BIOS to set it up, they essentially report it as being "dead" because it acts a lot like it. We did however figure out it's related to something very specific with the PCIe slots on the older BIOS version and the reason I didn't find this on testing is because I'd set it up, get IPMI working, and basically have to functioning well for the test, but once it reboots/IPMI gets reset, it still powers on/off, and we assume test was successful and send it out. Then DC gets it and can't boot into BIOS.

    At least one caught like this today and another caught maybe a few days ago. I'm hoping all the rest are the same and everything else I send out I'm going to update BIOS on.

    Only problem is it seems like if both BMC and BIOS are updated on 3000 series then it runs into problems. Just BIOS updated, seems to work fine.

  • @VirMach said:

    @tonyairicon said:

    @VirMach said:
    Fixed a bug with anti-abuse script which was causing some of the overloading/abuse problems by not catching them correctly on some Ryzen nodes.

    I did the migration and then the bandwidth seems to be having issues, 15000MB became 15000TB! As much as I would love to have that much bandwidth, it's probably really a bug.

    Description My migration payment ticket is Ticket #548489, you can check the bug according to this

    Bandwidth 0 B of 15000 TB Used / 15000 TB Free

    Looks like either we give 1GB bandwidth or 15,000TB. Nothing in between! We'll have it corrected but we're not too worried about it going this way, it's not like anyone is going to use 15,000TB. Don't prove me wrong.

    ok disturbed

  • FrankZFrankZ Veteran

    @VirMach - Yes exactly. I could never have said it so well.

    @AlwaysSkint RE: IIRC Marriott data was always coded with dates that included four digit years so Y2K was not really too big of a issue. Last time I was involved was 2002 and it was still going.

    I have this fussy recollection of using token ring with Novell NetWare back around 1987 while setting up a banking system on IBM XT PCs. I expect I have been trying to block it out ever since.

    Thanked by 2AlwaysSkint skorous
  • wenlongwenlong Member

    JP still not ready?
    My order #656149 still in Pending
    How long should I wait, is there a time schedule?

  • VirMachVirMach Member, Patron Provider

    @wenlong said:
    JP still not ready?
    My order #656149 still in Pending
    How long should I wait, is there a time schedule?

    UPS delayed package we sent last week on Monday, due to Japan holiday. It's arriving this Friday (today) and then if datacenter can set it up over the weekend it'll be done then, or most likely Monday.

  • JabJabJabJab Member
    edited May 2022

    Everything okey there with Ryzen Alpha aka Beta Node in PHX? ;)

    You probably also want to take a look at DENZ001 too.

  • VirMachVirMach Member, Patron Provider

    @JabJab said:
    Everything okey there with Ryzen Alpha aka Beta Node in PHX? ;)

    You probably also want to take a look at DENZ001 too.

    153 days uptime, RIP.

  • qpasnhqpasnh Member

    @VirMach said:

    @wenlong said:
    JP still not ready?
    My order #656149 still in Pending
    How long should I wait, is there a time schedule?

    UPS delayed package we sent last week on Monday, due to Japan holiday. It's arriving this Friday (today) and then if datacenter can set it up over the weekend it'll be done then, or most likely Monday.

    Can the TYOC030 network be repaired?
    The network seems to be broken drop

  • fanfan Veteran

    Just received an email about migration. It looks like my 64mb idler will get the first real migration among all. B)

  • VirMachVirMach Member, Patron Provider

    @JabJab said: You probably also want to take a look at DENZ001 too.

    Is that graph supposed to look like that? Had to rub my eyes and wonder if I should get some rest.

  • VirMachVirMach Member, Patron Provider

    @qpasnh said:

    @VirMach said:

    @wenlong said:
    JP still not ready?
    My order #656149 still in Pending
    How long should I wait, is there a time schedule?

    UPS delayed package we sent last week on Monday, due to Japan holiday. It's arriving this Friday (today) and then if datacenter can set it up over the weekend it'll be done then, or most likely Monday.

    Can the TYOC030 network be repaired?
    The network seems to be broken drop

    Yeah doesn't look good, not sure what happened, maybe the settings I applied dropped off. Checking.

  • VirMachVirMach Member, Patron Provider

    @qpasnh said: Can the TYOC030 network be repaired?

    The network seems to be broken drop

    Oh, the settings were not made persistent on a reboot and either it was rebooted or crashed briefly. Loading the settings back in.

    I'll also make these persistent soon.

  • JabJabJabJab Member

    @VirMach said: Is that graph supposed to look like that? Had to rub my eyes and wonder if I should get some rest.

    That is standard smokeping graph. From VPS IP to VNC Host IP (aka localhost outside VPS, no idea how to name that correctly :D).
    Black color shows 'variation' in probes timing - it was working nicely for weeks, then last two days seems to be fluctuating a lot (from 0.3 ms stable and 0 ping loss to like 10% ping loss and 0.3 - 5 ms ping return time).

    ping to Virmach.com https://i.imgur.com/h6YY73E.png
    ping to gateway https://i.imgur.com/DzwxOAu.png

  • VirMachVirMach Member, Patron Provider
    edited May 2022

    @qpasnh said: Can the > TYOC030 network be repaired?
    The network seems to be broken drop

    Really weird, I applied back the exact settings as before they got reset and it didn't look too great. Then I went back to completely different settings we used to use before I fine-tuned it to the final settings and it looks much better now.

    Looks like a little bit of RNG at this point but whatever.

    Old settings are still not perfect though but new ones aren't working, I'll leave it like this for now until I look into it further.

  • randomqrandomq Member

    @FrankZ back when I was a kid I networked a hotel's management/sales office so they could have file shares and email. Used 10base2 since the cards were cheap and no switch required. Unfortunately there was an oddball PC that didn't have the right drivers and could have ruined the whole upgrade, so they got a free upgrade from win3.11 to win95 at 2am.

    Hacked my high schools Lantastic token ring network, and we used a "hypercard" type application to make a virtual version of our school with jokes about all the teachers. Our own private metaverse in the 90's :)

    Thanked by 2skorous FrankZ
  • @VirMach Hi, I chose to migrate to Los Angles in the Ryzen button yesterday, and I read in your email just now that I will be migrated to the preference destination on Sunday. So does it mean LA nodes with Quadranet are online and ready to use?

  • MikePTMikePT Moderator, Patron Provider, Veteran
    edited May 2022

    @VirMach said:
    @LiliLabs just tagging you because we discovered something today relating to the X470 motherboards, so I thought you'd want to hear it. I said earlier this week that it was strange how all of them would test well, and then at the datacenter have problems to where DC hands couldn't even get IPMI set up, frozen at the screen.

    It turns out those are all X470 because of some bug in the BIOS. ASRock said not to update to the latest BIOS for 3000 series so I tried to listen... well, the old version apparently has this weird issue where on reboot when trying to get into BIOS it just freezes. If IPMI is set up, you can go in and use it just fine, but the problem is if they need to get into the BIOS to set it up, they essentially report it as being "dead" because it acts a lot like it. We did however figure out it's related to something very specific with the PCIe slots on the older BIOS version and the reason I didn't find this on testing is because I'd set it up, get IPMI working, and basically have to functioning well for the test, but once it reboots/IPMI gets reset, it still powers on/off, and we assume test was successful and send it out. Then DC gets it and can't boot into BIOS.

    At least one caught like this today and another caught maybe a few days ago. I'm hoping all the rest are the same and everything else I send out I'm going to update BIOS on.

    Only problem is it seems like if both BMC and BIOS are updated on 3000 series then it runs into problems. Just BIOS updated, seems to work fine.

    Glad to know that, I'm looking at purchasing the X470 version too, costs around 300€ here in Portugal.

    Why do AsRock say to avoid updating to the latest BIOS version?

  • VirMachVirMach Member, Patron Provider

    @JabJab said:

    @VirMach said: Is that graph supposed to look like that? Had to rub my eyes and wonder if I should get some rest.

    That is standard smokeping graph. From VPS IP to VNC Host IP (aka localhost outside VPS, no idea how to name that correctly :D).
    Black color shows 'variation' in probes timing - it was working nicely for weeks, then last two days seems to be fluctuating a lot (from 0.3 ms stable and 0 ping loss to like 10% ping loss and 0.3 - 5 ms ping return time).

    ping to Virmach.com https://i.imgur.com/h6YY73E.png
    ping to gateway https://i.imgur.com/DzwxOAu.png

    Focusing on PHXZ002 right now, doesn't look too great. IPMI is bugged out, waiting on DC hands.

    It just came back online on our monitor, but as IPv6. Also it looks like it never actually went "down" based on uptime, but I can't get into the server. Also it looks like all the VM processes are still running. But it's not responding on SolusVM. And neither are VM IPv4's. But monitoring timestamps are all correct and traffic is continuing to flow to VMs as normal. Except now monitoring is showing it offline on one page and online on another... and oh look we got one ping back, 2700ms 75% loss.

    In any case, has to be rebooted because of IPMI, wonder if it's somehow related and IPMI just gives a random unrelated error if it's having connectivity issues.

    And looks like PHXZ001 has the same problem too now but much less severe, oh and PHXZ002 is back up on SolusVM but still not responding.

    Yeah okay I think I just got thrown off by the IPMI being weird, otherwise it's just Phoenix connectivity issue.

  • @tototo said:
    Great tutorial by @FrankZ (I edited it a bit)

    To install O/S with CPU pass-thru enabled from a template:

    • From the billing panel, not SolusVM
    • Install template AlmaLinux8, wait 15 min or more,
    • then install the template you really want to install.

    To install O/S with CPU pass-thru enabled from a ISO:

    • From the billing panel, not SolusVM
    • Install template AlmaLinux8, wait 15 min or more,
    • then change boot order to cd-hard drive, mount ISO of choice, reboot,
    • install from ISO using TightVNC or similar VNC program,
    • unmount ISO, change boot order to hard drive only, reboot.

    How to find VNC login info (please scroll down!)
    https://billing.virmach.com/serverstatus.php

    If you don't want to change boot order/ISO mount, or want to use the latest netboot.xyz, try this.
    https://lowendtalk.com/discussion/comment/3406196/#Comment_3406196

    In all cases, check your boot order/ISO mount settings!

    (edit) This does not solve the hardware issue, but I post this because in many cases this helps.

    Unfortunately in AMSD026 node not possible this. After mount ISO the VPS didn't start. :(
    I could install Rocky Linux only tricky: first installed CentOS 7 with Reinstall-OS template and updated the CentOS 7.9 to Rocky 8.5 with http://itzgeek.com/how-tos/linux/centos-how-tos/how-to-upgrade-centos-7-to-rocky-linux-8.html.
    And if you want resize/shrink the templated partition/filesystem: http://unix.stackexchange.com/questions/226872/how-to-shrink-root-filesystem-without-booting-a-livecd/227318#227318.

  • pddpdd Member

    How long will the 3$ paid migration(without data)take effect?

  • VirMachVirMach Member, Patron Provider

    @pdd said:
    How long will the 3$ paid migration(without data)take effect?

    Is yours stuck?

  • VirMachVirMach Member, Patron Provider

    @schneitomee said: Unfortunately in AMSD026 node not possible this. After mount ISO the VPS didn't start.

    I'll resync ISO on these.

    Thanked by 1schneitomee
  • pddpdd Member

    @VirMach said:

    @pdd said:
    How long will the 3$ paid migration(without data)take effect?

    Is yours stuck?

    It seems so, about 20 days ago I tried to migrate to Tokyo but just one more useless ip and the ryzen migrate button disappeared, so today I tried again to migrate to Seattle for 3$ and again nothing happen! Here is the ticket

This discussion has been closed.