New GL-iNet hardware GL-MT2500 incoming

It is good then R4S?

What do you mean?
From what I know MT7981 (on MT2500) is 2 x A53, but RK3399 (on R4S) is 2 x A72 + 4 x A53 which is a lot faster. The MT2500 is just having one 2.5G port which doesn't quite make sense since it has no WiFi (unlike the MT3000 with WiFi 6), the other end using 1G port will make it not useful, that's why I am saying R4S is a lot better here.

I have a MT2500 that I've been waiting on for vanilla OpenWrt support.

My plan is to use the 1GbE port for WAN and the 2.5GbE port with a 2.5GbE switch for the LAN (and an AP off that if I need it, don't really use WiFi though). The MT2500 can sit right beside my modem, the switch will be in my panel box area where my home's Cat5e runs terminate, and the AP can be in any room of the house with a Cat5e run.

To me, that's an ideal, super-simple setup for 2.5GbE LAN with powerful WireGuard capability. For the foreseeable future I will not under any circumstances have internet faster than (or even close to, really) 1 Gbps, so it's absolutely perfect.

Then why do you need a L3 2.5gpbs device? A simple L2 (switch) with 2.5g ports would suffice

+1

if you're not hosting anything on the MT2500, requiring 2.5GbE, why bother ?
it won't make your lan "faster" by simply sitting there.

I sold my Brume 2 to a friend.
After installing the current master snapshot, only 494.30 MiB of disk space is available.
And after each restart the settings are reset and the installed packets are deleted.
Currently the mt2500 is not usable for my friend.

root@OpenWrt:~# df -h
Filesystem                Size      Used Available Use% Mounted on
/dev/root                 3.5M      3.5M         0 100% /rom
tmpfs                   494.3M      1.2M    493.1M   0% /tmp
tmpfs                   494.3M      3.0M    491.3M   1% /tmp/root
overlayfs:/tmp/root     494.3M      3.0M    491.3M   1% /
tmpfs                   512.0K         0    512.0K   0% /dev
root@OpenWrt:~#

Not sure how much sense it makes, but a single 2.5G port can be useful for a router-on-a-stick design, where you use a single link to connect the router to a switch with more than one L2 domain. Traffic routed between those domains will then traverse the link twice. Having 2.5G available means that it is possible to route full wirespeed between two 1G switch ports using such a configuration.

1 Like

The only useful case I can think of is VLAN routing, however I really doubt this CPU can handle this well given the limited processing power.

Did you installed wrong image?

I'm also willing to try on my 2500, but where is the right image?
I'd ill don't get any answer in nearest hours, I'll test the one from David, but please correct me if there's any "better" image to try.

No, I installed the sysupgrade image according to the instructions.

https://downloads.openwrt.org/snapshots/targets/mediatek/filogic/openwrt-mediatek-filogic-glinet_gl-mt2500-squashfs-sysupgrade.bin

i think i found the problem. mkfs.f2fs: not found

Sat Dec  2 20:57:44 2023 kern.info kernel: [   11.609817] loop0: detected capacity change from 0 to 15186944
Sat Dec  2 20:57:44 2023 daemon.info mount_root: overlay filesystem in /dev/loop0 has not been formatted yet
Sat Dec  2 20:57:44 2023 daemon.notice procd: /etc/rc.d/S95done: sh: mkfs.f2fs: not found
Sat Dec  2 20:57:44 2023 daemon.err mount_root: no jffs2 marker found
Sat Dec  2 20:57:44 2023 kern.info kernel: [   11.662319] loop0: detected capacity change from 15186944 to 15179776
Sat Dec  2 20:57:44 2023 kern.err kernel: [   11.671932] MTD: Attempt to mount non-MTD device "/dev/loop0"
Sat Dec  2 20:57:44 2023 daemon.err mount_root: failed - mount -t jffs2 /dev/loop0 /rom/overlay: Invalid argument

Installed and it booted. Unfortunately I am using this device for multi wan via 3 USB's and that's currently broken until mwan3 is correctly rewritten using nftables. I tried compiling with full support and editing boot, but it worked in a weird fashion. I jumped back to last compile from glinet infrabuilder, where you could compile a plain openwrt build.

1 Like

Anyone know why this missed the 23.05.3 release in March, did it fail to build?

Support was added to main in Nov. 23 and never backported to the 23.05 branch.

1 Like

Will this device be covered by the next stable release? I think I will wait for a stable release

3 Likes

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.