primeval_god
-
Posts
869 -
Joined
-
Last visited
-
Days Won
2
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Report Comments posted by primeval_god
-
-
3 hours ago, jaclas said:
I'm probably not looking where I need to. Please show a screenshot if you have it configured.
If you select the "file descriptor limit" graph labeled fds in the side bar, then one of the dimensions on the graph should be shfs. I use an older version of the ui so i just annotated your screen shot with where to look.
-
5 hours ago, jaclas said:
Thx, I did it, restarted the container.... and now where to look for this value in netdata webgui?
Under the Applications section, the graphs should now have a shfs line. For this specific issue the subsection apps.files will give you the number of open files for the shfs process.
-
17 hours ago, jaclas said:
@primeval_god I have installed netdata, but I have never used it and it is completely unknown to me. I'm looking for how to create the configuration you write about, but I can't find anything sensible.
Can you help? Any link? The netdata documentation is an ocean.
You will need to obtain a copy of the apps_groups.conf file from the netdata container. You can use the docker cp command for this. From the command line, in the appdata folder where you store your netdata configuration, run
docker cp netdata:/etc/netdata/orig/apps_groups.conf apps_groups.confassuming your container is named netdata. Add the line
shfs: shfsto the end of the apps_groups.conf file. Then all you need to do is bind mount the file back into the container at the path /etc/netdata/apps_groups.conf (add the entry to the docker template, make sure both the host and container paths are to the file itself).
-
On 5/2/2024 at 12:47 PM, jaclas said:
I also came across an interesting thread on reddit, where someone stated that the problem is due to a very low limit of open files. That could be right, because in the background my server was processing thousands of mp3 files and I was programming in Node.js at the time, where a project in the ./node_modules subdirectory can also have tens of thousands of files.
If you happen to use Netdata for system monitoring you may be able to spot the circumstances leading up to the crash if the open file limit is the issue. Netdata has a section under Applications (apps.files) that shows the open files for different process groups. By default SHFS is not listed, but its open files are likely group under system or other. With a little customization of the netdata config you can make shfs show up as a separate entity. (Creating a custom apps_groups.conf file in the container and adding line like "shfs: shfs")
-
JorgeB dug up this piece of info.
-
1
-
1
-
-
8 minutes ago, JorgeB said:
No, it's automatic, though you could create a folder with the same name of the share in the array if you wanted to disabled it for some reason, then re-start array.
That is not ideal. The reason is that I have no need for IO performance beyond what shfs already provides. I would much rather have the functionality that is mentioned under the exclusive share restrictions (in particular the ability to make changes and add folders to other disks without having to stop and start the array).
-
Does the exclusive share behavior apply automatically to shares with Primary->Pool and Array->None or is it an option? Is there an option for a Primary->Pool and Array->None share to not have the exclusive share behavior?
-
Just want to bump this issue as, based on the discussion here https://forums.unraid.net/topic/134119-btrfs-on-unraid-checksum-on-read-when-transfering-data-to-windows-pc/?do=findComment&comment=1220283,
it appears to still be relevant in the 6.12 beta.
-
The root of this problem was the "patch ui" setting of the compose manager plugin. Among other things it patches it changes the naming scheme of the icons and is apparently buggy on 6.11.5.
-
Interesting so there may be a bug with the compose plugin gui patch on 6.11
-
On 1/1/2023 at 2:21 PM, gmight said:
I had one container enabled using compose. I've deleted it since it didn't give me the desired outcome (a while back). It was a gluetunVPN (2). But like I said, it's gone.
The others are all built using unraid templates.
Is the compose manager plugin installed? If it is check its settings page and make sure the "Patch UI" setting in disabled.
-
On 10/8/2022 at 12:21 PM, mgutt said:
If you directly close the browser, it will run forever, too.
Maybe what the webui needs is a sort of reaper task to periodically clean up the nchan processes if no clients remain connected. Can nchan detect if there are active subscribers to a channel? If not is there a way for a nchan client page to send a periodic keep alive back to the server?
-
1
-
-
I know this report is quite old but was this issue ever investigated or reproduced in 6.10+?
-
I love the idea of having a locked release announcement thread! I would like to see this paradigm extended to the threads in the Announcements category of the forum as well, though maybe with a link to a discussion thread in that case (see here). This makes it much easier to get notifications when something is announced or updated rather than whenever someone posts a comment about the announcement.
-
47 minutes ago, jonathanm said:
I'm not sure what you are referring to, as Unraid hasn't ever released a version with nvidia drivers.
If you are referring to using the community modded version of 6.8.3, I would contend that using the official 6.9 beta is far more "production" ready than using a community mod.
In principal I agree, however though I am not using it myself I think it unfair not to acknowledge the stability attained by @CHBMB and @bass_rock 's Nvidia plugin. From its forum thread, it appears that there are quite a number of successful users and few to none show-stopping issues. That seems pretty stable to me.
Here is hoping that the limetech's latest addition will prove itself as stable in short accord.
-
1
-
[6.8.3] shfs error results in lost /mnt/user
in Stable Releases
Posted
That shouldnt be a problem. Perhaps its a permissions issue? Do the permissions on the apps_groups.conf file match those of the other files in /etc/netdata? I assume you restarted the container after adding the config file, but if not that should be done as well.