Horaco 2.5GbE Managed Switch (8 x 2.5GbE + 1 10Gb SFP+)

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

DRUROB

New Member
Jun 26, 2026
4
0
1
But this is the unmanaged version right? the webinterface wouldnt be available incl the hidden menu´s?
 

Shonk

Active Member
Nov 25, 2016
260
185
43
75
e.g. mine with 1.9.1 firmware is
SWTG118AS-V2.1-17462 with a datecode of 2023-10-21
 

Shonk

Active Member
Nov 25, 2016
260
185
43
75
if i doesnt work let me know i will temp put my spi dump up later
but make sure you change your mac as i dont want duplicates of my mac really
 

DRUROB

New Member
Jun 26, 2026
4
0
1
Now i know! ok, i found the label on the board and i have the same ...exactly the same datecode :)

but it still has the blinking sys led and i cant connect to the switch .
i would highly appreciate the firmware dump of your switch
 

Shonk

Active Member
Nov 25, 2016
260
185
43
75
make sure you tick verify some ch341's are ropey when connected to a board
i have like 5 and only two are reliable when on a board
 
  • Like
Reactions: DRUROB

Seth_space

New Member
Aug 10, 2025
5
1
3
Great suggestion that already implemented in the last release 2026.5.1 where i added many features
I am using v2026.6.5 and its a nice addition to my homelab. Thanks a lot!
The scanner option does not seem to see any device. I see the activity in the logfile but no results in the scanner screen.
 

McArrow74

New Member
Nov 28, 2025
25
16
3
I am using v2026.6.5 and its a nice addition to my homelab. Thanks a lot!
The scanner option does not seem to see any device. I see the activity in the logfile but no results in the scanner screen.
I'm using v2026.7.1 and the scanner seems to work very well. I have about 5 Horaco switches and the switch dasboard is really cool...
 

Attachments

Seth_space

New Member
Aug 10, 2025
5
1
3

dp8208

New Member
Jul 13, 2026
1
0
1
Hi, i have bricked Sodola SWTGW0108P , PCB v1.04, does anyone can send me a complete dump to recover? I have programmer and tried many fw files, but nothing works...


bricked2.jpg
 

Takeshi

New Member
Jul 18, 2026
1
2
1
After struggling some days with a bricked KeepLink KP-9000-9XHPML-X-AC (PoE version) I got it back to work. I had to replace the flash chip after a leg broke. Flashing to a 4MB chip I stole from a broken ESP8266 ended in continuous reboots without error message. Apparently, the firmware checks the UID of the flash chip and resets on a mismatch.

I have updated the firmware to skip this check and it now works perfectly.

Device ModelKP-9000-9XHPML-X-AC

Firmware VersionV100.9.5
Firmware DateFeb 25 2025
Hardware VersionV1.1

I have also figured out the u-boot password for those interested:
Qa3e7ARx#

and here's the patched firmware skipping the UID check:

Full dump:

Runtime (web) upgrade:
 
  • Like
Reactions: RTurala and blunden

jomama22

New Member
Jul 16, 2026
2
1
3
When my KP9000 was dead, this dump actually ran it perfectly, and the web interface was completely different — it identified itself as SWTG218AS. The only problem was that I couldn’t log in because I didn’t have the password. Can anyone figure out what this is actually used for?
Way late for you but on my xikestor 4x 2.5 2x sfp (not sure why the firmware is named the same, maybe you used the wrong one, anyway...) you need to connect to serial, start it up and login with uboot password: Switch321 (this is through serial, not web).

Then reset user config with 'reset' command and then 'reboot' command to reboot. Unplug everything and then start as normal. 192.168.2.1 is the login page, admin admin is login.
 

peppebytes

New Member
Apr 23, 2026
9
14
3
Thanks.. will look for the updat then. Pitty its not in the Docker Hub Container Image Library | App Containerization directory. That would update easy

edit: github only has a 2026.6.4
Follow the instruction here you can intall the latest release as the release page is only a description with a package of the last zip of the repo that you can get with clone.
 

Seth_space

New Member
Aug 10, 2025
5
1
3
Follow the instruction here you can intall the latest release as the release page is only a description with a package of the last zip of the repo that you can get with clone.
Thanks, did install the more recent version but still the scanner function does not see anything.
Looking at the logs its starting and scraping the switches mac table but to no avail.
"HC-SWTGW218AS"

there are mac entrys in the dashboard page but the scanner shows 0
 
Last edited:

Escobar69PL

New Member
Aug 21, 2026
3
3
3
Hi @Takeshi and Everyone,

I have a very interesting case with two KeepLINK KP-9000-9XHPML-X-AC switches (the 8x 2.5G PoE+ / 1x 10G SFP+ version) and I wanted to share my findings and full dumps with the community.

Both of my switches have the exact same motherboard revision: 2M-PCB51-ML-V1.1 (based on RTL8373N + RTL8224N). However, they came with different factory firmwares and behaved completely differently.

- Switch #1 came with firmware V100.9.25 and it works perfectly fine and stable with my 4K PoE cameras.
- Switch #2 came with a weird factory firmware v100.9.9.1.7 and it was extremely unstable (randomly losing passwords and dropping PoE power).

I contacted KeepLINK support regarding Switch #2. They sent me an official firmware file: "runtime_9XHMPL-X_keeplink_v100.9.14.bin" (the M and P in the filename were swapped by their engineers, but it's 100% for the PoE version). I flashed it to Switch #2 via WebUI and it successfully fixed the password reset bug. Interestingly, the KeepLINK representative claimed that version V100.9.25 does not officially exist in their database!

Since they didn't know about V100.9.25, I disassembled my stable Switch #1 and used a CH341A programmer with a SOIC8 clip to read the SPI Flash directly. The flash chip used on this 2M-PCB51-ML-V1.1 board is a 2MB capacity (detected as W25Q16). The reading process was 100% successful and verified.

I have attached the clean 2MB full dump of the elusive V100.9.25 firmware to this post (zipped).

Could anyone, especially @Takeshi, help with patching this V100.9.25 dump to bypass the UID verification check? I would love to convert it into a safe web-upgrade file so I can upgrade my second switch to V100.9.25 as well.

P.S. If anyone in the community needs the official v100.9.14 web-upgrade bin file that I received from KeepLINK support, please let me know and I will gladly upload it here too!

Thanks for your amazing work and community support!
 

Attachments

RTurala

New Member
Jun 20, 2020
1
0
1
After struggling some days with a bricked KeepLink KP-9000-9XHPML-X-AC (PoE version) I got it back to work. I had to replace the flash chip after a leg broke. Flashing to a 4MB chip I stole from a broken ESP8266 ended in continuous reboots without error message. Apparently, the firmware checks the UID of the flash chip and resets on a mismatch.

I have updated the firmware to skip this check and it now works perfectly.

Device ModelKP-9000-9XHPML-X-AC

Firmware VersionV100.9.5
Firmware DateFeb 25 2025
Hardware VersionV1.1

I have also figured out the u-boot password for those interested:
Qa3e7ARx#

and here's the patched firmware skipping the UID check:

Full dump:

Runtime (web) upgrade:
Could you write how have you disabled UID check?
 

pkoci

Member
Aug 13, 2023
32
2
8
Hi,


has anyone encountered packet loss with a Raspberry Pi 5 (SATA SSD Samsung 870 powered via USB, and a USB ZBT-1 dongle) running the latest version of HAOS? I have an unmanaged Horaco 8-port switch with 1x SFP+. I tried different ports and cables without any luck. I then connected the Pi directly to my Mikrotik hAP ax3 router and the packet loss seemed to disappear — though that test only ran for a few minutes, so I can’t be fully sure the switch was the cause.


Since then I’ve noticed something that might be a stronger lead: I have an add-on running a continuous 1-second ping to the router in the background. Every time I stop it, packet loss comes back within seconds and stays elevated for hours (survives reboots too). Starting it again brings loss back to near zero. This has happened twice, in both directions, but I’m not 100% sure it’s the actual cause rather than a coincidence — could still be something on the switch side too. Interface error counters (rx_errors/dropped/crc) stay at zero throughout either way.


So I’m not fully ruling out the switch, but I’m now leaning more toward something on the Pi’s onboard NIC / link-state side (possibly EEE-related, unconfirmed). Curious if anyone with a similar switch has seen packet loss on a Pi 5 (or other gigabit-only device) specifically, and whether constant traffic on the link made a difference for them too.