Unable to Update Fresh Install

SoCalPilot

Member
Sep 13, 2022
20
1
8
Just installed ProxMe VE 8.4-1 on my 45Drives HL15 server and ensured the ISO has the proper SHA256SUM after download. After first login I went to the update section, disabled the enterprise repo, enabled the no-pve repo and then tried to perform an initial update however received a bunch of "Hash Sum mismatch" errors and was unable to update.

I rebooted the server then tried `apt clean` then `apt update` and `apt upgrade` but receive same “Hash Sum mismatch” errors.

I edited /etc/apt/apt.conf.d/99fixbadproxy and added...
Acquire::http::Pipeline-Depth 0;

Acquire::http::No-Cache true;

Acquire::BrokenProxy true;

Acquire::Retries "3";
Then tried `apt clean` and `apt update` and `apt upgrade` and this crashes the entire server, forcing a manual reboot.

I tried `apt clean` then `apt update` and `apt upgrade` again but it continues to give the same errors.

(I tried this on two separate NVMe drives and also ran MemTest86+ for 20hrs for 4.5 full passes and found zero errors.)

Not sure what happened or what other steps to take here. Any tips or advice?




Server Specs:
45Drives HL15 2.0 Server
Motherboard: ASROCK ROMED8-2T
CPU: AMD EPYC 7452 (Cores: 32 Threads: 64)
Memory: 256GB (Qty 4 - Micron 64GB DDR4-3200 ECC Registered RDIMM)
HBA Card: LSI 9400-16i
Boot Drive: Kingston NV2 1TB M.2
Add-on Card: Supermicro AOC-SLG3-2M2-O (Qty 2)
NVMe Pool: Samsung 990 Pro 4TB NVMe (Qty 4 - ZFS RaidZ1)
HDD Pool: Seagate Exos 24TB (Qty 5 - ZFS Raid Z2)
GPU: Nvidia GeForce RTX 4070 Super
IPMI: ASPEED AST2500 BMC
 
Since it was a fresh install I tried the "Helper Script" to install a Ubuntu VM and it froze on the download then eventually gave an error...

✔️ https://cloud-images.ubuntu.com/noble/current/noble-server-cloudimg-amd64.img
############ 16.2%curl: (56) Recv failure: Connection timed out
[ERROR] in line 467: exit code 56: while executing command curl -f#SL -o "$(basename "$URL")" "$URL"


I tried other OS's and all seem to give the same error making me feel the problem is on my side, but its a fresh install and Im not behind any kind of proxy. Just Fiber coming into a Ubiquity Router & Switch, directly to the server.

... When I click on the link above it downloads to my desktop without any issues and at a fast speed.
 
Last edited:
Hi. In G. search results it seems that a reason happens to be some some device in the network (e.g. a switch in between or even some other device in the network).

Can you completely, galvanically cut of all devices (leave only the server and the router)?
 
If you are not behind a proxy, you don't need the proxy configuration.
Since it's a fresh install, did you test basic networking and dns ?
Also why not try PVE 9.0 for a fresh install ?

I did try PVE 9 and was having a bunch of similiar yet random issues so I thought maybe going to 8.4-1 would yield more stable results... Which doesn't seem to be the case. Im having a ton of network issues for some reason but am having a problem nailing it down even on fresh installs. I believe my ISP (AT&T Fiber) does not support IPv6.
 
I tried proxmox-ve_9.0-1.iso again (Freshly downloaded the iso and verified checksum) and then reloaded it onto my 45Drives HL15 server through IPMI. Installation went smooth. Logged into the server for the very first time after the installation was completed and before I started tweaking any settings I went to download an ISO that I know to be good from a known good source.

I immediately had issues...

downloading http://xxxx.com/bnc/Win11_25H2_English_x64.iso to /var/lib/vz/template/iso/Win11_25H2_English_x64.iso
--2025-10-20 16:48:00-- http://xxxx.com/bnc/Win11_25H2_English_x64.iso
Resolving xxxx.com (xxxx.com)... 64.91.238.197
Connecting to xxxx.com (xxxx.com)|64.91.238.197|:80... connected.
HTTP request sent, awaiting response... 302 Found
Location: https://xxxx.com/bnc/Win11_25H2_English_x64.iso [following]
--2025-10-20 16:48:01-- https://xxxx.com/bnc/Win11_25H2_English_x64.iso
Connecting to xxxx.com (xxxx.com)|64.91.238.197|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 7736125440 (7.2G) [application/octet-stream]
Saving to: '/var/lib/vz/template/iso/Win11_25H2_English_x64.iso.tmp_dwnl.2960'
0K ........ ........ ........ ........ 0% 22.3M 5m29s
32768K ........ ........ ........ ........ 0% 32.7M 4m36s
65536K ........ ........ ........ ........ 1% 32.8M 4m17s
98304K ........ ........ ........ ........ 1% 32.8M 4m7s
131072K ........ ........ ........ ........ 2% 32.6M 4m1s
163840K ........ ........ ........ ........ 2% 33.0M 3m56s
196608K ........ ........ ........ ........ 3% 32.9M 3m53s
229376K ........ ........ ....... 3% 32.7M=8.0s
2025-10-20 16:48:09 (30.9 MB/s) - Read error at byte 259343891/7736125440 (Decryption has failed.). Retrying.
--2025-10-20 16:48:10-- (try: 2) https://xxxx.com/bnc/Win11_25H2_English_x64.iso
Connecting to xxxx.com (xxxx.com)|64.91.238.197|:443... connected.
HTTP request sent, awaiting response... 206 Partial Content
Length: 7736125440 (7.2G), 7476781549 (7.0G) remaining [application/octet-stream]
Saving to: '/var/lib/vz/template/iso/Win11_25H2_English_x64.iso.tmp_dwnl.2960'
[ skipping 229376K ]
229376K ,,,,,,,, ,,,,,,,, ,,,,,,,. ........ 3% 45.9M 9m32s
262144K ........ ........ ........ ........ 3% 45.1M 4m5s
294912K ........ ........ ........ ........ 4% 43.3M 3m28s
327680K ........ ........ ........ ........ 4% 45.5M 3m11s
360448K ........ ........ ........ ..... 5% 43.9M=3.5s
2025-10-20 16:48:14 (38.1 MB/s) - Read error at byte 400423715/7736125440 (Decryption has failed.). Retrying.
--2025-10-20 16:48:16-- (try: 3) https://xxxx.com/bnc/Win11_25H2_English_x64.iso
Connecting to xxxx.com (xxxx.com)|64.91.238.197|:443... connected.
HTTP request sent, awaiting response... 206 Partial Content
Length: 7736125440 (7.2G), 7335701725 (6.8G) remaining [application/octet-stream]
Saving to: '/var/lib/vz/template/iso/Win11_25H2_English_x64.iso.tmp_dwnl.2960'
[ skipping 360448K ]
360448K ,,,,,,,, ,,,,,,,, ,,,,,,,, ,,,,,... 5% 49.0M 35m48s
393216K ........ ........ ........ ........ 5% 26.4M 6m20s

But what's even stranger is that it crashed my entire ProxMox server, here are the kernel logs...
https://pastebin.com/BXwxmfYJ

For reference this is a brand new NVMe drive I purchased today (Samsung 990 Pro) because prior to this with the issues I was having I thought maybe my NVMe drive was bad and figured a replacement would do the trick. Also worth nothing, I ran MemTest86 for 20+ hours and completed 4.5 test with 0 errors detected.

This is literally a brand new ProxMox installation, I touched zero settings. Just went to download and ISO from a known good source and that failed, then crashed the whole server. Im a bit confused on what to debug here at this point.
 
Rebooting the server, and trying to download another ISO from a known good source immediately crashed the server again, pretty much the exact same errors as last time. This is on a fresh install, zero configuration changes made.

Rebooted and trying to let the system install the listed available packages and that fails also, same errors Ive been having across all previous installations...
https://pastebin.com/akbMSecJ
 
Last edited:
Have you checked dmesg and jornalctl?
Run memtest? Disk check? Updated all firmware?

Yes, with the help of ChatGPT I checked dmsg and jornalct logs which always pointed to either some kind of faulty hardware or memory issue. Server is brand new and I even put in a brand new hard drive today thinking it was a failing drive, none of that helped. I ran memtest86 for 20hrs, 4.5 full tests and it reported zero errors.

I do like the idea of trying to install Ubuntu, Debian or heck even Windows OS to see if that somehow works, thus helping me to rule out a hardware issue.
 
Update: The other day while trying to troubleshoot another error, I updated my bios to the latest firmware 4.10. I just randomly had the idea to go back to what I had before which was 3.9 and Im immediately noticing a huge difference. Finally I was able to update my system with zero errors and now downloading ISOs seem to be working, thus a HUGE improvement for me. I continue to test.

So for those with the ASROCK ROMED8-2T motherboard, consider NOT upgrading to Bios 4.10 and remain on 3.90. I hope this helps someone else.