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.
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
This discussion has been closed.
Comments
You should be fine in Phoenix.
@Virmach I presume that you realise SAEZ001 is still as dead as a can of spam.
Btw, any chance PHXZ002 have XPG drive?
great! hopefully we could be included here finally
Yeah, got the DC to finally move the disk over but they fumbled it a bit, trying to figure out where it is right now. If I can't get data off then we'll just recreate these as well and ship the disk over for a further look.
I have a list of all XPG drives right now, PHXZ002 does not have an XPG drive.
My support ticket has been closed, do I still need to send you a new support ticket?
Ticket for what, TYOC038? There's a network status open, we're still aware, and we already provided an update that you quoted. No ticket is necessary for that.
TYOC039 update we were able to get the disk back up. I'm not booting any virtual servers so we can get a quick transfer off for these VMs. It shouldn't take long.
List of servers with probably XPG, due to the way the PCIe risers work in these I can't be 100% sure until going into BIOS but it's a high likelihood:
SEAZ002
SEAZ003
PHXZ001
ATLZ005
These definitely have an XPG:
TYOC040 (2)
TYOC039 (And it's broken, people are getting moved right now.)
SEAZ001 (And it's broken.)
TYOC038 (And it already broken and people got moved.)
TYOC037
TYOC035 (And it's already broken and people got moved.)
TYOC034
These are broken in some way or missing disk and may be XPG, but no one will be affected, but it's likely it's completely unrelated.
SJCZ009
PHXZ003
All others confirmed to not have XPG or already gotten rid of it.
These are transferring off at an average 500MB/s. Should only be another half hour of downtime for TYOC039. IP addresses will stay the same.
@VirMach I should get the hynix p31 if i buy more flash storage on my pc right?
Overall, yes, I'd say it's the best drive. In terms of price, latency, and performance once it's full and where it matters.
TYOC039 took a little bit longer as it ran into some I/O errors during the transfer but it's about to be done and everything is going to be booted back up.
zug zug
@VirMach The last Ticket
#794490
Chicago update. Psychz says this is ready now, and we need to send in the "correct" PDU to match the incorrect setup they did, so I've ordered one off Amazon to arrive Tuesday. We'll send off the servers as well and hope they do the right thing on Amsterdam. Of course I'll still keep an eye out for communication on Monday on Amsterdam.
SEAZ001 update. DC hands and the team are struggling to understand my request. Disk was moved to potentially the wrong server which is fine but none of the servers detect it and they won't confirm where they put it. Last real information I got is that there's no 4TB drive, only 8TB (they're referring to an HDD. For some reason they can't tell the difference between PCIe and HDD.)
I'll try to provide more updates later. Still waiting on SEAZ001 response and in the meantime packing while I do a few other things I talked about earlier today.
This is why we can't have nice things.
I went over these numbers a while ago, but I just looked them up as a refresher. Something like 40 days time to kill a drive per 500GB and that's consumer grade/average. So, close to 3 months and $129 at Hetzner to kill two 500GB drives. Except that's based on one old source that doesn't provide much information, and if Hetzner is using Samsung, a 250GB drive lasts around 5000TBW. Even if we assume the 500GB is only as good as the 250GB, let's say that's 2,500~ plots at 6 hours a plot. That's closer to 625 days for a single drive and 1250 days for two of them. Hetzner makes $1,763 before their $100 in disks are killed.
Maybe my math is completely wrong, I just remember looking all this up more thoroughly in the past and walking away from it without any concerns for losing money on dedi abuse.
None nvme samsung dies till now for me and I am on 12month now for 500GB, the writes where 100% 24h now slowing/stoping as dedi will e used for other stuff. but all 10+ drives health is 100% (no spare block used zero issue withing smart data) even the drives should be dead long time ago at least it stand for samsung SAMSUNG MZVLB512HBJQ
critical_warning : 0x4
temperature : 53 C
available_spare : 100%
available_spare_threshold : 10%
percentage_used : 255%
data_units_read : 9,336,543,944
data_units_written : 13,229,512,524
so that one is perfect health and TBW on 500GB model as atm 6160TB
I don't know what that title is about, but a lot of these TBW tests test well beyond real world failure points and useless for such comparisons. I don't give a shit how long for the whole drive to get wrecked, it's unstable and blue screening after the first blocks get retired. Amplification writes goes to the fucking moon. The warranty also long blown, and newer models don't use MLC.
Your calculation also for some reason only takes revenue into account for the drives and not for the entire machine, so erasing the two year roi for the whole server isn't seen. In other words, if they expected to make ~$100 profit in 625 days, they won't any longer.
Lastly, they had to take action against this due to failed drives and the increase in technicians swapping drives and customers delayed/impacted. Try and argue with real world at your own peril.
Please look work order #749044. TYOC039 VPS suddenly lost its hard disk and stayed offline of a day long.
Will the remaining Tokyo March pre-order vps run on the existing nodes or will they run on the upcoming new storage nodes?
Please look what VirMach posted a while back!!
The only point I was trying to make is that it's likely that the impact is likely less than what was circulating at the peak of the anti-Chia movement and I of course over-simplified it to represent it as a potential ratio of the total revenue, which is often what companies may use in their pricing formulas when compared to the cost to acquire the hardware.
Of course you do make a good point, when going in more depth, the negative externalities are more than just the cost of the drives. I definitely did not include that portion.
Are you asking if the Ryzen VPS advertised with NVMe run on the HDD storage node?
Hi, first of all please do not use the outage report button like that. You didn't add the node name to the title, so it's not helpful, and it would just be closed. Also, we already sent out emails, as well as updated network status page for this issue.
And we've also already fixed these as well.
Yours is not directly related to this issue anymore. It seems like after the disk issue/outage, you spammed the buttons, some within the same minute, like re-install, boot, reboot, rescue mode, boot, re-install, then a couple minutes later re-install, boot, boot, re-install, etc. This caused yours to fully break and disappear. I'd recommend closing the outage report and use the appropriate button to create a ticket and explain what you did in what circumstance and mention that your disk is fully missing and it needs to be recreated.
@VirMach My VPS was disabled because the traffic was wrongly set to 1G, please check
I believe you heard yesterday that the issue has already been recognized by VirMach and is awaiting a fix by their developers.
Please sync some ISO on ATLZ005 if you have free time on the next days :-)
Thanks.