adrianedelen

Sorry for the delay, got caught up with work. After restarting so far it seems like everything is running normally. Attached are the updated diagnostics. Do you know what may be happening that would require a reboot. normally I am able to leave the server on indefinitely, usually only shutting down when I need to do maintenance, or the power will be out. gary-diagnostics-20230421-1740.zip

Recently (around 1 week ago) I noticed my docker image started to grow. After some reading I found typical causes to be writing files to the image instead of to a mapped path, so I started to look into that route. I however wasn't able to find anything. Later I decided to change the docker image size since it was approaching 100% and I didn't want to fill it up ( I hadn't found the root cause of the growing image size, however I also did not want everything to go offline as I needed some services available) after changing the docker image size (increasing) the docker service failed to start throwing these PHP errors Warning: stream_socket_client(): unable to connect to unix:///var/run/docker.sock (No such file or directory) in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 712 Couldn't create socket: [2] No such file or directory Warning: Invalid argument supplied for foreach() in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 898 ok, I have seen mention of this on the forums before, seems like a corrupted docker image. no big deal, I deleted the image, rebuilt it, and this time I didn't install about 1/2 of the previous dockers that were not as important in an attempt to reduce the image size and find the offending container that was causing the image to grow. after getting everything going again, I logged in this morning only to find websites failing, and containers not responding. I saw the container was showing ~85% usage of allocated space (30GB) so I went to the docker page and calculated the image size, unfortunately I didn't take a screenshot, so I don't remember the size for the logs/writable portions, however I know the total container size was showing as 30.5GB. Something isn't adding up, the dashboard is reporting ~85% but the calculated size is well over the allocated size. I shut down the docker service, went to change the image size again to increase to around 60GB (the previous size I had used for years with no problems) and now the docker service is failing to start again with the same PHP errors. I saw that issues like this can arise from full drives, but everything is reporting with 100s of GB of free space on every drive, with the cache drive have ~500GB free. In summary: Docker image growing Resized image Docker service failed to start rebuilt image with less containers docker image filled up again resized image docker service failed to start again. Thanks for any help! gary-diagnostics-20230421-1032.zip

I run a lot of non-priority containers on my network that will occasionally demand 100% or the majority of system resources. an example would be cloud backups that will occasionally read from several disks and consume a high amount of bandwidth. This is fine except for scenarios when I need those resources actively, for example plex. A typical scenario may be: 1. Watching a movie on plex 2. another container starts an operation that is indexing files and uploading them. 3. this causes plex to buffer. I have tried to schedule operations to occur during offpeak hours, but it can be hard to manage, since the time I or someone else would be using plex is not consistent. So my question is: I know that CPU priority can be configured, Is it possible to give the Plex (or any given specific docker) container higher priority for disk access and network bandwidth?

After making the cstate change it looks like server performance has improved. If we go another day or two without a crash I will consider it resolved and mark the solution. Thanks for the help.

Ok I am looking into these changes now. I will try those out and report back on the next crash. In the meantime, I had a crash (lockup) around midnight last night. Attached is the syslog as it is one the flash drive. it seems to be enlightening, but a little beyond my comprehension syslog

Attached is the diagnostics zip. I will update this post with the syslog after the next crash (I just restarted so it may be a few hours) Thanks, for the quick response. bignas-diagnostics-20230224-2031.zip

I have been experiencing periodic crashes as of late, at differing times, it seems to happen no later than 24 hours after a boot up, the server will become entirely unresponsive, even in gui mode, the screen is locked up, no mouse/keyboard input. I can attach log files, I am just not sure what to attach. I am at a loss as to what is causing this and could use some troubleshooting guidance. Thanks.

I figured out the problem. Mods can close this topic if needed. For anyone in the future, it was a pretty dumb problem. For some reason with the new hardware the static IP for unRAID was wiped (locally) I booted into safe mode, no plugins, was able to connect to webui locally and re assign the IP address.

I recently upgraded effectively my entire server. Motherboard, cpu, and ram, also added an additional drive. At the moment the server is not connected to any network (still setting up other changes), but I am able to log in to the local gui, however when attempting to connect to the web interface, Firefox responds with "Unable to connect, an error occurred during a connection to 192.167.1.42:81" The port is correct. The same happens when using localhost, and the local url.

Thanks for that suggestion, I will check this out. I will do a test run tonight, using my current setup. What I am trying to avoid is commiting to swapping all of this hardware around, getting the rackmount case, etc. if it is not a suitable setup.

Lately I have been considering the idea of swapping my server with my desktop PC. My server is an old surplus lenovo s60 with an ancient xeon, not enough sata, and slow ram. My current desktop is much more powerful, and seeing as I am using my server to do more and more heavy lifting, I figured it would make sense to swap (or just get rid of the server and get a cheap desktop). My intention/goal would be to install/transfer unraid and drives to desktop PC. and then create an appropriate VM on unraid for my needs. Then on my new cheap desktop, just connect to the Vm on unraid. Kind of like a thin client. Additionally, My server is about 40 feet away with several walls in-between and it is connected via cat6 and various switches. I am still new to using VMs and I currently have a few running on unraid, I connect to them on my desktop using TigerVNC, however, there is a lot of latency, stuttering mouse, artifacting, and other nuisances that would make using the VM as a daily machine an unpleasant experience. I am willing to sacrifice the overhead of virtualization, But I am wondering, is there a better way of connection to a VM then something like vnc? is daily driving a VM on a remote machine, with gpu passthrough a reasonable thing to do? Additionally: I have considered using SFP/in wall fiber for things like usb/display port/etc. however this is not Ideal as i would like the felxibility of ethernet (and running that stuff through the attic is a pain)