dlandon

The Open Files plugin is working fine with unRAID 6.3.2.

Version 2017.02.19 has been released. I have learned that inotify is pretty chatty when a lot of disk activity is occurring and the Files Activity log file can get excessively large. I have released a new version that will keep the log file to the last 20,000 events so it doesn't fill the log space.

Glad to hear you are using it. Someone mentioned that there wasn't a posting for quite a while and they thought the plugin might have been abandoned or dead. Not so. I too like making my own choices for the colors. I have a color deficiency and some of the colors I have trouble discerning.

It appears that you are using rsync in the script. UD has no control over what rsync is doing. You might find an rsync parameter that would respect the special characters in your file names or a way to display them in their original form.

Just to let everyone know this plugin is still alive and well. It just hasn't needed any updates, but is still working with unRAID 6.3.

The Recycle Bin plugin manages samba deleted files and the trash you are talking about is created with NFS mounted shares. I don't think it's a good idea to try to manage NFS trash with this plugin. This would accomplish what you want. Once a day? Good idea guys. Yup, this would be perfect You would have to create a script to do this. I would suggest using User Scripts and set it up to run once a day.

What are the original file names?

I get it. The reason I added that to the title is because I have gotten tired of the constant "My disks are spinning up for no reason. Cache_dirs must be broken." Hopefully this plugin can show people what disk activity is happening and they can understand why disks spin up.

The 'appdata' share is always excluded just for the reason you mentioned - too much activity and there is no value in tracking that.

New version available 2017.02.18a allows you to enable/disable UD and cache. I don't like the idea of automatically excluding SSD disks. There may be a reason someone wants to see disk activity on a SSD disk. This way the user can make the decision on which ones to include/exclude.

The Recycle Bin plugin manages samba deleted files and the trash you are talking about is created with NFS mounted shares. I don't think it's a good idea to try to manage NFS trash with this plugin. This would accomplish what you want. Once a day?

Which parts? EDIT: So I'm thinking: UD Disks Cache As the choices?

I have worked out how to log open read and write file events so the plugin will log read and write file activity. It turns out that inotify is excessively chatty when tracking open file and directory events. I managed to filter out all the extraneous open events that don't apply to disk spin ups. Hopefully now all file activity that can cause disk spin ups will be logged. I'll release a new version tomorrow after a bit more testing.

Version 2016.02.16 displays many of these messages on the console: find: '/tmp/.preclear/': No such file or directory

There is a new version of the file activity plugin. I changed to using inotify to log the write disk activity.

Please provide more information. I do not understand why this is a problem. Samba handles the delete function. Why is Sonarr managing the .Recycle.Bin? How does it even know of its existence?

If you have installed this plugin, please use the following link to update the plugin. Because of a bug in my plugin, it will not update through the plugins update. Remove the previous version first. https://github.com/dlandon/file.activity/raw/master/file.activity.plg This will load the latest version that will update properly.

Linux has an audit function that I am looking into that is a log of all file activity. The problem is that the 'audit' functionality has to be built into the kernel. Without it the 'audit' package I need to use won't work. I will need LT support on enabling the 'audit' capability. I'm not sure exactly where to go with this right now. What I am trying to do is answer the age old question of why disks are spinning up and why doesn't cache_dirs prevent it (it can't on writes). There is a lot of mis-understanding of what cache_dirs can and can't do, and some unrealistic expectations. I am hoping to provide a tool to show users what is accessing disks and making them spin up. The Open Files plugin shows currently open files and which processes have them open. The Files Activity plugin shows the write history within the last two hours on the disk, cache, and UD shares. Very different animals.

The plugin has to query the disks to determine the last file and directory modify times. There is currently no way to store this information in a log. Linux has an audit ability to show all file activity, but the kernel has to be built with that ability enabled. It may be worth the effort for me to investigate that further. I am trying to help answer the questions about disks spinning up.

Not sure. They kind of do different things and it might be confusing to combine them.

This plugin shows files on the array that have been written or modified recently. Open files shows currently open files. These could be open for read or write access.

This is a plugin that shows the file and directory activity on the array by disk. This includes read and write activity on the array. Read activity can cause disk spin ups. It is recommended that you have the Dynamix Cache Directories plugin installed and running. The plugin uses inotify to log read and write disk activity. You will need to start File Activity to start the logging. I don't recommend leaving it on all the time. A server with a lot of file activity can fill the log space. You can see which files and directories are being read and written and keeping the disks from spinning down. The file activity is shown by disk, UD disks, and the cache and Pool drives. The 'appdata', 'syslogs', and 'system' shares are excluded. This is not the ultimate answer for disk spin ups, but I used it to find an issue with Windows 10 backing up files every hour. It would write new files on the cache, but constantly modify the older file directory entries causing spin ups on disks. The plugin is installed in the Tools menu under System Information. The preferred way to install the plugin is with CA. Or you can install the plugin by copying the following URL in the install plugin line of unRAID: https://github.com/dlandon/file.activity/raw/master/file.activity.plg

Once the current jobs I'm running are completed on my VM's, I'll reboot. Was trying to avoid the reboot if possible, but seems like it will be unavoidable. Thanks again! Reboot worked. Thanks for responses, and thanks for pointing out how to do an NTFS checkdisk. I was not aware of that functionality. Is there a chance you removed the device while it was still mounted? All of the different file systems on UD devices can be checked in the same way.

I don't think there is anything for repairing an NTFS filesystem on unRAID. You will have to plug it into a Windows computer and let it checkdisk. UD can check the NTFS file system. Click on the + next to the mount point when the disk is not mounted and you can click on the graphic to check the file system. You are given an opportunity to do a repair if needed. Hover your mouse over the active items and you'll see a tool tip describing what the action will be if the active item is clicked.

Reboot.