thestraycat Posted January 3, 2018 Share Posted January 3, 2018 Hi guys, I've noticed that whenever i'm fooling around in the unraid gui whilst watching a film from my unraid array via LibreElec (XMBC) The client buffers when im navigating the menus etc... has anyone else experienced this issue? Anyone know where to start troubleshooting it? I've recently moved house and have set up the network again from scratch, blatted the router config, remove devices from the network, removed complexity of having multiple switches to the point that i now simply just have unraid attached via 1 cable to my switch attached to my router. Everything is cabled is attached to the switch which is currently just unraid and my HTPC the rest is on wifi. My issues are with devices that are cabled however. Any ideas? Quote Link to comment
Frank1940 Posted January 3, 2018 Share Posted January 3, 2018 You might want to clarify your configuration a bit. I assume that you have a free standing HTPC to feed the video to 'Movie' display setup. Now what you are using to browse the GUI. I see that you have a Xeon processor which is a bit of overkill if you are running a simple NAS setup. But I would bet that you are not. Tell us what else you are doing with all that Xeon horsepower. You might also tell us if the freezing occurs on DVD type material, BluRay or both. You should also attach the diagnostics file which might provide a clue if there is something going on with the server... Quote Link to comment
thestraycat Posted January 4, 2018 Author Share Posted January 4, 2018 (edited) Totally. Sorry i just didn't want to put people off by writing a monster thread So yes it's a Xeon, but it's an E3... first gen. A glorified i7 really. I run around 20 containers mainly idling and doing nothing. SABNZBD is my heaviest hitter and i havnt yet found a nice way of utilizing the Nice IO parameters and multipar settings to alleviate stress on the array when a large download comes down which is defintely a problem when sabnzbd is processing a large download however i am running some NiceIO and multpar settings but they dont seem to really be doing anything - However my issue occurs when sabnzbd is idle so i'm removing that from the equation right now. My Unraid box is usually idling around 12gb RAM usage out of 32gb and <10% across all cores when watching a film on my HTPC. My HTPC is an older i5 NUC cabled up with gigabit (the same as my unraid NAS) and the issue occurs when natively playing a 300mb tv series episode or a 50gb 4k dump. I'm browsing the unraid GUI via chrome on a macbook via WIFI. The weirdness results in the native movie file 'buffering' on my HTPC as the file is played natively over an NFS share and a gigabit network this simply means that the array has either stopped responding and the movie pauses until the array resumes again as oposed to any transcoding happening from a plex server. I've even disabled docker and run the NAS just as an array to see if ANY of the containers are conflicting and the issue still happens. To me, it feels like when i'm working on the array maybe extracting some data, moving a folder etc that i see buffering. I have also simply been logged into the Unraid GUI clicking about an experienced the issue when nothing has been happening on the array. It nearly feels like some background service, check, inner working of unraid is locking the array ocassionally but that's merely a guess. I have a 10gb nic that i plan on installing and using connected to my gigabit network just so i can remove the variable that it's the nic, although i dont think it is as the logs show nothing of note happening with the NIC. Edited January 4, 2018 by thestraycat Quote Link to comment
pwm Posted January 4, 2018 Share Posted January 4, 2018 Have you verified the network connectivity with your unRAID machine at times when everything freezes - i.e. can you ping it? Quote Link to comment
thestraycat Posted January 4, 2018 Author Share Posted January 4, 2018 its only happens for around 5-8 seconds and it happens randomly... so tough to do.. .I have try to hit the array via chrome and it just spins until the array starts responding again by which time the film resumes. Quote Link to comment
pwm Posted January 4, 2018 Share Posted January 4, 2018 Open a console somewhere and have it ping continuously. Then it's enough to quickly look at the console when you see a freeze. Quote Link to comment
thestraycat Posted January 4, 2018 Author Share Posted January 4, 2018 true. i have thought about it. I'll keep it running on a VM. IF there are drops what would you conclude? Quote Link to comment
pwm Posted January 4, 2018 Share Posted January 4, 2018 If ping continues to work, then it's likely you have load spikes, out-of-memory issues or similar in unRAID. If ping stops for a couple of seconds then it's more likely something happens with the networking somewhere. Getting info from ping is just a way to try to figure out in which direction to continue searching for the problem. Quote Link to comment
thestraycat Posted January 5, 2018 Author Share Posted January 5, 2018 Makes sense. I'll get one running. Quote Link to comment
thestraycat Posted January 9, 2018 Author Share Posted January 9, 2018 Interestingly enough PWM i've just noticed that the "fix common problems plugin" has recently mentioned i'm suffering from out of memory errors... If i upload logs any chance you can cast an eye? Quote Link to comment
Frank1940 Posted January 9, 2018 Share Posted January 9, 2018 20 minutes ago, thestraycat said: "fix common problems plugin" has recently mentioned i'm suffering from out of memory errors This has fixed the problem fro a lot of folks: https://lime-technology.com/forums/topic/58855-regular-out-of-memory-problems/#comment-577424 Quote Link to comment
thestraycat Posted January 9, 2018 Author Share Posted January 9, 2018 I'll take a look. I've had a look at glances and memory dosn't seem to be a problem. Do we know what triggers the need for the fix mentioned? Glances: Quote Link to comment
Frank1940 Posted January 9, 2018 Share Posted January 9, 2018 I Think that the OS allots 20% of available as a continuous block of memory at boot time. In your case, this is 6.4GB of RAM. As the system runs, memory becomes more and more fragmented. If a larger block of memory is requested then is available in a single block, you have an out of memory condition. unRAID can't roll an unused process out to the cache disk because it doesn't use a cache disk. So it resorts to other means to get the memory which usually involves killing some process. Now what is unRAID using this block of memory for? Well, it is to implement a concept called 'write delay' which was intended to make a system always seem responsive to user input (like typing) when a disk write as called for by some process on multi-tasking system. The"write"would be sent to this block of RAM (Reporting back to the requesting process that the write was finished) and doing the actual write to the physical disk when the CPU was idle. (By the way, this was introduced back in the Windows 3.1 or Win95 days when the 80360 processorand 1MB of RAM was a hot item!) (Note that with one megabyte of RAM, this is only 200KB not the 6.4GB in your system!) Many of us feel that this function is not really very applicable today. Modern HD have very large RAM buffers on them to do the same sort of thing and the CPU's are so much faster that any delay due to a disk write operation is inconsequential. You won't hurt anything by trying it and it has fixed the issue for many folks. Quote Link to comment
thestraycat Posted January 11, 2018 Author Share Posted January 11, 2018 Thanks Frank1940 - i agree. I'll try the fix. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.