Nothing happens when I type any thing. Normally if I reset the box, I can type stuff on screen.
The ZC is not accessible via ssh nor going to the IP on the browser and I don’t see it on the zima client
I didn’t do another screenshotas what you saw on the last one is what happens. Nothing changes or feedback, same screen stays on, it doesn’t matter what you press
If I were to restart the device by doing a hard-reset, then everything works, I can type and connect to the Zimacube, but a few hours of inactivity, this happens, it doesn’t respond, and it’s not visible on the network
But the hard drives are spinning, I cnd hear them and the Zima cube lights are on, the 7th bay lights come on and off, just the system is completely frozen
I have 6 4TB Western Digital red Plus hard drives and 1 SAMSUNG 970 EVO Plus SSD 1TB NVMe M.2
Configured as two RAID 5s (3-4TB HDD each) and the nvme just by itself for now
I think we have located the cause of the problem. Before the next version is released, you can also help us verify the accuracy of the solution with simple commands, if you are willing:
We first guessed that it might be related to the temperature increase or CPU usage, but after deploying the monitor scripts on the test environment equipment with abnormalities, we found that there was no sign of high CPU usage and temperature increase even in the last second before each restart, so we denied our guess.
Later, we found the clue to the problem that [kswapd0] was stuck several times when the CPU usage was high, and no operation could be performed, and it could not be connected through LAN, but the fan was still running and the CPU was still hot. By querying the information, it was found that it might be related to the configuration of swappiness, and then confirmed that there was actually no swap partition inside the system. By the way, other suspicious configuration items were checked and the configuration problem of RAM overcommit was found.
The first command is to turn off the use of swap partition. The purpose of swap partition is to use the disk as virtual RAM when our RAM is not large enough. But its limitation is also obvious, that is, the read and write speed of the disk is not as fast as RAM, and the gap is very large, which may be hundreds of times. Since we do not prepare a swap partition due to system limitations, it is appropriate to turn off this configuration, otherwise it will hang or restart after being triggered because there is no available space in the swap partition or there is no swap partition.
The second command is to turn off overcommit of RAM. This is actually used in conjunction with the swap partition. After it is turned off, overcommit of RAM will no longer be supported.
OK, this is getting to be a massive headache. I got a few days reprieve where my Zimacube stayed on the network and I thought perhaps this had been fixed. I was wrong. Last two days my cube has gone offline even though it had not lost power and was still on and running when I went to reset it. This is a major inconvenience for me since my router and cube are upstairs in my office. I am 67 yrs old and have a difficult time going up and down the stairs and this guarantees that I will have to go up and reset it each morning. NOT ACCEPTABLE!
The whole reason to own a NAS is for continuous access from all the devices on my network. I run Plex on another computer and all my media files disappear when my cube goes down. NO WAY I would even consider moving the server to my cube right now. Which again, is part of why I bought it in the first place.
My son has a Synology NAS and he has always questioned my decision to buy my Zima instead of an established product. Now he laughs at me when I tell him my cube is down AGAIN. I am also trying to be active answering questions about my cube on Facebook and social media but it is getting more difficult to stay positive regarding my answers.
This is clearly an issue that multiple users are having and yet it has not been addressed. You need to get this dealt with. I’m losing confidence in your ability to correct such a basic issue as staying connected to a network.
I tried it and it did not work, the system still froze,
Just to clarify, the CPU usage should be low, as this seems to happen when the system is not in use for a couple of hours. Something may be happening when it just idles for a long time. It freezes and is unreachable via IP and the OS as seen on the screenshots is just frozen, I can’t type anything until I restart the ZC again.
Hi Michael, I’m sorry for the very inconvenience. If you have time in the next few days, could you provide a remote environment to help us further investigate? In order to investigate this issue and verify the effectiveness of the solution, we have verified that at least two abnormal devices have not had the previous frequent restart or freeze problem for nearly a week. We may need to confirm whether the solution is effective and check whether there are other mixed scenarios that cause restarts or freezes.
Hello, naz5442, thank you very much for your assistance in verification. If it is convenient, can you provide remote information to help us confirm the environment and troubleshoot possible further problems again? Your observation and feelings are correct. Many times, the freeze occurs during periods of low usage.
Yes i should be able to do that. In reading the other users, mine usually happens overnight so also during low usage periods. I was able to see several days in a row, nearly a week, with no issues but all of a sudden the last couple of days it reoccurred.
For whatever reason, my Cube has stayed online and functional since my last post.
Still, the one change I made was disconnecting ZimaClient on my iMac and MacBook. It’s not running on my MacBook, but I noticed it’s still running on my iMac, it just isn’t connected.
If the crash is from a RAM/swap issue, could it be the part of ZimaOS that interfaces with the client? Disabling the client seemed to be the main workaround that has helped a couple people. However, there was a message in another thread where it didn’t help the person. And still, I had one freeze several days ago without the client.
From the information I have collected, this problem is indeed complex. The client is one reason, but as you said, it is not the only reason. The general scenario of encountering this kind of problem is on a Cube with many apps installed, even if it is not used or idle.
I noticed that many of the apps are running in RAM and CPU.
Something else I noticed, not sure if it’s related to the issue, is when I had a virtual machine running, the virtual machine did not show up in either CPU or RAM usage. The gauges increased, but not the processes with their percentage / RAM amount.
The above investigation has been completed. Directly disabling the swap partition cannot solve the problem in many scenarios. We are considering replacing the new solution to solve the current problem by adding a default swap partition while ensuring the availability in high memory usage scenarios. We conducted actual tests in some environments provided by community users and found that when there are apps with high memory usage, the availability of the entire system will be affected.
FYI i am also still having issues with mine. It’s not every single night but probably 4x per week at least i have to unplug my cube and restart it before my network can see it. Power light is still on when i have to unplug it. Keep us posted please.
It started happening to me again yesterday. Twice in a couple hours.
The trigger that appeared to cause it was booting up my rarely-used PC, where I had installed ZimaClient and ZeroTier some time ago. This computer is usually turned off.
I don’t think ZeroTier is the issue as it’s still running on my two Macs, and I’ve used it to connect to the Cube remotely a few times without issue.