Update to OpenWrt to Cudy X6
- Posted in:
- OpenWrt
Here is an update on Cudy X6 after close to two months of use.
After I made Cudy X6 my main router, I had 2 cases where I shortly lost connectivity for about 10 seconds. First time I was quite busy and then forgot it, but the second one happened when I had time to research and I found this in logs:
NETDEV WATCHDOG: eth0 (mtk_soc_eth): transmit queue 0 timed out
…
[572627.361199] mtk_soc_eth 1e100000.ethernet eth0: transmit timed out
[572627.370851] mtk_soc_eth 1e100000.ethernet eth0: Link is Down
[572627.421492] mtk_soc_eth 1e100000.ethernet eth0: configuring for fixed/rgmii link mode
[572627.429576] mtk_soc_eth 1e100000.ethernet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
This issue was happening around once per 10 days. After some research I decided to disable “Hardware flow offloading” and it looks like it fixed this issue. To my surprise, I found that it didn’t affect the speed at all. I think this router has enough CPU power to transfer data fast enough without hardware offloading.
The router works fine for about a month but yesterday I found that 3 days ago my router was restarted. It was restarted 10 minutes after midnight and usually, we are sleeping for a couple of hours. As result, it does not look like somebody did it. So, there are the following probabilities:
- There was a power outage. I checked and all other devices weren’t restarted. It could be a very short one that affected only this device and didn’t affect other devices.
- The router crashed and restarted. I investigated this and it does not look like it is the case because in this case crash handler is supposed to create a log file but it is not there.
- There was a hardware issue that lead to a reboot without the router knowing about that. For example, bad power supply, overheating CPU, etc.
- I messed up something in the configuration.
- Some bug in OpenWrt
I researched case #5 and I couldn’t find anything. It is too general and as a result, I had to discard it.
I will keep monitoring the router and if this will happen again then most likely it is case #2 or #3. For case #2 I’m planning to create a permanent kernel log that writes directly to flash memory. If I don’t see anything the next time it happened, then it means that most likely it is case #3. I will do a factory reset and set it up again just to be safe to rule out case #4.
One restart per month is not bad at all and there are not many routers with the original firmware that can work that long, but I would like to have at least a few months of uptime.