@VirMach said: I believe all the Ryzen templates are on passthrough now as of like yesterday actually.
I've been running some re-installs on AMS Test Server AMSD030X/RYZE.AMS-D001.VMS ..
Has WHMCS/SolusVM been messing you about again? Alama 8.3/Debian 10/Ubuntu 20.04(yuk!) all give Epyc Milan.
(No selectable ISO in SolusVM.)
Just saying.
Network has epic speeds (geddit?).
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
The no ISO might just be because of the new plan I made for it and forgot to include the ISO group but I actually don't remember if it's using its own or NVMe1G. Could just mean NVMe1G doesn't have it either and no one's noticed.
Milan thing, I have a guess but not really sure. Possibly related to the version of libvirt/qemu we have on it. I assume it was added in when Milan came out at some point and they just mis-identify it? If that could be a thing? I just personally can't wait until people start throwing out conspiracy theories about how we're giving Milan instead of Ryzen to screw customers over or something like that.
Does it actually result in lower Geekbench score/affect the clockrate?
@VirMach said: Does it actually result in lower Geekbench score/affect the clockrate?
Highly unlikely.
In other news: to add to my year of misery, I broke another tooth yesterday and, after pulling out the cracked bit today, the dentist says its' remnants gotta get extracted. Can't be done until I get off the blood thinners for a while. Jeez!
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
@VirMach said: Does it actually result in lower Geekbench score/affect the clockrate?
From what I have seen in the past the clock rate is unaffected, but the Geekbench score is slightly lower. This is not a new thing and I have seen it happen before on the old E5's where it would show the CPU as a E3 with a clock speed of 2200 mhz. On the old servers installing a debian 6 template would almost always fix it.
If you wanted to confirm that it is a libvirt/qemu version issue maybe you could use virsh edit to manually change the cpu mode in the VM config file to <cpu mode='host-passthrough'/> and see if it fixes the issue. If it did then you would know it was an issue with SolusVM and not the libvirt/qemu version.
@AlwaysSkint said: In other news: to add to my year of misery, I broke another tooth yesterday and, after pulling out the cracked bit today, the dentist says its' remnants gotta get extracted. Can't be done until I get off the blood thinners for a while. Jeez!
I used to have these kinds of problems, but do to a "situation" I was involved in some years ago I found my self without any teeth. Afterward I had titanium implants and cubic zirconia crowns installed as replacements. Best thing I ever did as these things are indestructible. I highly recommend them.
@FrankZ said:
I had titanium implants and cubic zirconia crowns installed as replacements. Best thing I ever did as these things are
indestructible. I highly recommend them.
@VirMach said:
Milan thing, I have a guess but not really sure. Possibly related to the version of libvirt/qemu we have on it. I assume it was added in when Milan came out at some point and they just mis-identify it? If that could be a thing? I just personally can't wait until people start throwing out conspiracy theories about how we're giving Milan instead of Ryzen to screw customers over or something like that.
Does it actually result in lower Geekbench score/affect the clockrate?
Doesn't seem to, single threaded GB5 performance is within 5-10% of one with a 5900x, close enough I'd call that a margin of error considering various other people on the node. Clock speed matches the known 5900x one as well, so I'm going to assume that's a 5900x in there. I was mostly just curious which cpu the node had since I keep track of all of that kinda stuff in a spreadsheet of my services, but I got my answer haha. I think the bigger concern for people is you can't get VM-x/AMD-V enabled with any of the templates on this node either, though it's not something I really need.
Yeah it's not looking very good for FFME004. Two DIMMs aren't showing up, but are causing a ton of problems. I can't get logs to work so I can't really do much in terms of seeing what happens. And to top it off, an NVMe SSD also simultaneously failed and isn't coming back.
If I had to guess based on what I'm seeing so far it might actually just be the board and the DIMMs and NVMe are fine.
@gogogo said:
Hi @VirMach , i want to migrate my vps from NYC to LA, But I don't find the migrate button, what should I do?
Best case scenario it means we don't wish to sell any additional services to you such as migrations, most likely due to not following instructions in the past, creating too many unnecessary tickets, etc. If your service was originally in another region we'll add a button in the future to where you can request to move back to that region.
Other possible worse reasons:
Multiple accounts
Chargebacks, severe abuse, or illegal activity on your account
Being involved with force migration to Tokyo, account selling/trading, or BF special sniping
You can give me an invoice ID if you want me to check the specific reason.
We were able to bring FFME004 back online, but it only has half the physical memory functioning.
This means that there is not enough RAM for all virtual servers at this time. To minimize downtime while we schedule hardware maintenance, we will immediately be migrating approximately half of all virtual servers to another nearby server. We will begin by moving the service with the highest memory usage to minimize the quantity of services migrated, so there's a higher likelihood of you being affected if you have a larger memory service.If your service is migrated, it will return online with a new IP address. You may view the new IP address on your service details page, and may have to use the "reconfigure network" button or manually configure the new IP address within the operating system, but in most cases this will be completed automatically.
While this migration is progressing, your virtual server may be temporarily shut down and virtual server controls will not be available. Please keep in mind if you are not migrated, there will be an upcoming maintenance window which we will notify you of in a separate email, when it's arranged with the datacenter. If the situation worsens, we may elect to do immediate emergency maintenance instead.
We apologize for the inconvenience. This round of maintenance/migration should be completed over the next 8 hours. We will post any additional updates on the network status page.
Will I be a lottery winner in this migration too? Let's see, I have 4GB server there
Thanks for fixing 04 with half of RAM (and SSD was boot?). MySQL replication managed to catch up with 8 hours downtime, I am impressed.
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
It appears that the issue that @bakageta described earlier is not limited to Node DALZ008. I just reinstalled from CentOS7 to AlmaLinux8 via template on Node NYCB015 and ran into the same issue where the CPU passthru and VM-x/AMD-V is not enabled.
@FrankZ said:
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
It appears that the issue that @bakageta described earlier is not limited to Node DALZ008. I just reinstalled from CentOS7 to AlmaLinux8 via template on Node NYCB015 and ran into the same issue where the CPU passthru and VM-x/AMD-V is not enabled.
@FrankZ said:
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
************************ via templates, *****************************
******************* use ISO like sane human ************************
@FrankZ said:
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
************************ via templates, *****************************
******************* use ISO like sane human ************************
There, fixed that for you.
Interesting point - while I completely understand your point e.g. it is the safest, you know what you have installed etc.
My point would also be - if the supplier can't make a template that works and is secure, then surely this shows up other much greater problems associated with the suppler.
Can / should we trust AWS templates, Vultr, DO etc?
I feel a working and secure template should be part of the contract for the VM. If passthrough is supposed to be provided it should work in the template etc. The supplier should say these are 'working / approved' templates (which they guarantee should run) and these are 'other' templates (which may run).
BTW this is not meant to reflect on Virmach - just a general point.
@FrankZ said:
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
************************ via templates, *****************************
******************* use ISO like sane human ************************
If you have that already disabled via templates then reinstalling via ISO won't help.
Don't get it disabled via templates in the first place and just reinstall via ISO while it's already (still?) enabled
Jesus, this is just PSA to PSA that you can still reinstall, just don't use templates atm because seems like they broke/n.
Sure, VirMach will fix it one day, but if you really really need to do and need that VM-x - it just do it via ISO rather than templates until they fix it.
@Jab said:
Don't get it disabled via templates in the first place and just reinstall via ISO while it's already (still?) enabled
Really only helpful if you're reinstalling, I don't believe new services will have cpu passthrough / vm-x enabled from the get-go, they certainly haven't been enabled on new services in the past.
Comments
I've been running some re-installs on AMS Test Server AMSD030X/RYZE.AMS-D001.VMS ..
Has WHMCS/SolusVM been messing you about again? Alama 8.3/Debian 10/Ubuntu 20.04(yuk!) all give Epyc Milan.
(No selectable ISO in SolusVM.)
Just saying.
Network has epic speeds (geddit?).
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
The no ISO might just be because of the new plan I made for it and forgot to include the ISO group but I actually don't remember if it's using its own or NVMe1G. Could just mean NVMe1G doesn't have it either and no one's noticed.
Milan thing, I have a guess but not really sure. Possibly related to the version of libvirt/qemu we have on it. I assume it was added in when Milan came out at some point and they just mis-identify it? If that could be a thing? I just personally can't wait until people start throwing out conspiracy theories about how we're giving Milan instead of Ryzen to screw customers over or something like that.
Does it actually result in lower Geekbench score/affect the clockrate?
Just can't get the staff, these days, huh?
LOL
Highly unlikely.
In other news: to add to my year of misery, I broke another tooth yesterday and, after pulling out the cracked bit today, the dentist says its' remnants gotta get extracted. Can't be done until I get off the blood thinners for a while. Jeez!
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
From what I have seen in the past the clock rate is unaffected, but the Geekbench score is slightly lower. This is not a new thing and I have seen it happen before on the old E5's where it would show the CPU as a E3 with a clock speed of 2200 mhz. On the old servers installing a debian 6 template would almost always fix it.
If you wanted to confirm that it is a libvirt/qemu version issue maybe you could use virsh edit to manually change the cpu mode in the VM config file to
<cpu mode='host-passthrough'/>and see if it fixes the issue. If it did then you would know it was an issue with SolusVM and not the libvirt/qemu version.I used to have these kinds of problems, but do to a "situation" I was involved in some years ago I found my self without any teeth. Afterward I had titanium implants and cubic zirconia crowns installed as replacements. Best thing I ever did as these things are indestructible. I highly recommend them.
For staff assistance or support issues please use the helpdesk ticket system at https://support.lowendspirit.com/index.php?a=add
Doesn't seem to, single threaded GB5 performance is within 5-10% of one with a 5900x, close enough I'd call that a margin of error considering various other people on the node. Clock speed matches the known 5900x one as well, so I'm going to assume that's a 5900x in there. I was mostly just curious which cpu the node had since I keep track of all of that kinda stuff in a spreadsheet of my services, but I got my answer haha. I think the bigger concern for people is you can't get VM-x/AMD-V enabled with any of the templates on this node either, though it's not something I really need.
could this be you , far into the future?
https://en.m.wikipedia.org/wiki/Crystal_skull
I bench YABS 24/7/365 unless it's a leap year.
Yup, missus has one. Don't think I'm worth the couple of grand needed though and mine are near the back. Have never been that fond of soup. Slurp.
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
@Virmach, FFME004 just went off, doing anything?
Yup, also FFME004 - this is different though.
I mean it's dead :P
Noticed at: 2022-10-19 02:41:37 (UTC+01:00)
Encountered errors:
Amsterdam: Timeout (5 sec)
London: Timeout (5 sec)
Frankfurt: Timeout (5 sec)
Warsaw: Timeout (5 sec)
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
Yeah it's not looking very good for FFME004. Two DIMMs aren't showing up, but are causing a ton of problems. I can't get logs to work so I can't really do much in terms of seeing what happens. And to top it off, an NVMe SSD also simultaneously failed and isn't coming back.
If I had to guess based on what I'm seeing so far it might actually just be the board and the DIMMs and NVMe are fine.
Hi @VirMach , i want to migrate my vps from NYC to LA, But I don't find the migrate button, what should I do?
Best case scenario it means we don't wish to sell any additional services to you such as migrations, most likely due to not following instructions in the past, creating too many unnecessary tickets, etc. If your service was originally in another region we'll add a button in the future to where you can request to move back to that region.
Other possible worse reasons:
You can give me an invoice ID if you want me to check the specific reason.
@VirMach --- HI
guess what i wanna ask about?
Uhhh... transfer ticket or free SSD2G?
Will I be a lottery winner in this migration too? Let's see, I have 4GB server there
Thanks for fixing 04 with half of RAM (and SSD was boot?). MySQL replication managed to catch up with 8 hours downtime, I am impressed.
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
correct .... i will ask again next month .
i know your busy just checking its still pending from your side
have a nice week ahead.
and don't forget you have a life also.
LMFTFY
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
@VirMach
SEA with 47.87.193.*，When you could solve the problem? Can not use for 2 weeks. Please give a relpy!
************************ NOTICE **********************************
To anyone that requires VM-x/AMD-V for nested virtualization to be enabled
********************* DO NOT REINSTALL ***************************
It appears that the issue that @bakageta described earlier is not limited to Node DALZ008. I just reinstalled from CentOS7 to AlmaLinux8 via template on Node NYCB015 and ran into the same issue where the CPU passthru and VM-x/AMD-V is not enabled.
*This has been a public service announcement by @FrankZ.
For staff assistance or support issues please use the helpdesk ticket system at https://support.lowendspirit.com/index.php?a=add
Oh sad news
There, fixed that for you.
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
Interesting point - while I completely understand your point e.g. it is the safest, you know what you have installed etc.
My point would also be - if the supplier can't make a template that works and is secure, then surely this shows up other much greater problems associated with the suppler.
Can / should we trust AWS templates, Vultr, DO etc?
I feel a working and secure template should be part of the contract for the VM. If passthrough is supposed to be provided it should work in the template etc. The supplier should say these are 'working / approved' templates (which they guarantee should run) and these are 'other' templates (which may run).
BTW this is not meant to reflect on Virmach - just a general point.
Ok, AlmaLinux9 via netboot ISO
For staff assistance or support issues please use the helpdesk ticket system at https://support.lowendspirit.com/index.php?a=add
If you have that already disabled via templates then reinstalling via ISO won't help.
Don't get it disabled via templates in the first place and just reinstall via ISO while it's already (still?) enabled
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
Jesus, this is just PSA to PSA that you can still reinstall, just don't use templates atm because seems like they broke/n.
Sure, VirMach will fix it one day, but if you really really need to do and need that VM-x - it just do it via ISO rather than templates until they fix it.
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
Really only helpful if you're reinstalling, I don't believe new services will have cpu passthrough / vm-x enabled from the get-go, they certainly haven't been enabled on new services in the past.
Is this one of the ASRock x470 boards? If you're not using a riser for nvme it's likely the board. The onboard nvme fails pretty often on those.