sota

Would like to note: Same issue for me after a restart after 131 days (trying to sort out another problem.) Immediately had a problem starting any containers. Found this thread, #commented out the noted line (for now), restarted. Containers are starting again. While i'm glad this fixed it, can I get some info as to what/why this happened? AKA, did I miss an announcement someplace? Thanks!

also i'm getting a lot of this crap, filling up the logs... s.sqlQuery QUERY ERRORED INSERT INTO `Timelapse Frames` (ke,mid,details,filename,size,time) VALUES (?,?,?,?,?,?) s.sqlQuery ERROR Error: ER_GET_ERRNO: Got error 194 "Tablespace is missing for a table" from storage engine InnoDB

Anyone have any suggestions for something "dumber" than shinobi? I've decided it's too tempermental for my needs. I just need to dump the live stream from multiple cameras, to separate folders, broken up in 5 minute increments. I can live without the 'timelapse' image option, as it's problematic at best, and I can create post processed. I don't need any of the fancy that shinobi provides, as I use the native app to my cameras to view them in real-time, or I go look at the MP4 files directly.

as long as it's branded HP, sure.

you're better off abandoning the onboard P410i and installing an H220 or H240 controller instead. you're going to hate life, trying to get drives to work right with the P-series controllers.

Someone can correct me if i'm wrong, but the only time parity isn't maintained is if the array is inactive, and you mount the disks outside of the array.

I'm still suffering with a timelapse problem. it's set to grab a frame every 6 seconds, but when the cameras go from daylight to night mode, it drops to 1 frame every 15 seconds. it didn't used to do this.

the switch is pretty painless as well.

Howdy. I'm trying to migrate from the brettm257/unifi container to this one. my current controller version shows up at 5.14.17, and I'd like to keep that for the time being. I can't find the appropriate tag to install that version to start with. any help picking the right one would be appreciated.

been quiet for a while. had an incident where everything shut down. seems like a couple forks won't re-sync for some reason. haven't cared enough to really dig into it at this point.

probably going to start version/revision updates in the next couple of days. that should be fun.

still at the 2XCH, and a monster amount of some of the sh!tcoins. farms haven't dropped any more drives lately. I'm not worried about them in the garage and the cold. in fact, they might keep the entire garage warmer this winter.

Let me start with, I am changing disks out frequently without powering down, and this "issue" has cropped up before, with no solution presented. I currently have a share that appears for a disk that's no longer connected to the system. Further, if I attempt to reconnect the disk, the logs will show something about duplicate UUID being present, and refuse to even show the drive in UD. I need to find out where UD stores attached disk info, in the hopes that I can edit something that'll make the phantom share/drive disappear, so it can be reattached, without having to reboot the box. Thanks!

*sigh* looks like 2 more disks on farm1 and 1 disk on farm2 might be in trouble.

and since i'm not planning on doing any more plotting, guess I'm just going to have to live with 830 instead of 866. at least for the time being.

well, shit. just lost a drive in farm1. 36 plots, poof.

aye the issue is the decoding/extracting frames takes a brutally long time. each storage drive is: 4 cameras (currently. more to come hopefully) 13 days (roughly) 24/7 recording means I need to extract 60*60*24*13*4=4,492,800 frames total. then recombine them back into an MP4 @ 30fps. I'm fine with leaving the encoding part as CPU based, as that's usually running about 15fps, and once I have all the extracted frames I can easily batch process that to run. That'd take 83.2 hours, or just about 3.5 days continuous, to process. I'm fine with that part. it's the extracting that takes over a week to do. Yea I know shinobi does timelapse, but it sucks at 6 seconds per frame, so it's easy to miss things going by the cameras. And I want the 1fps files so I can delete the "useless" footage, as I go through and tag only the time stamps that have interesting things going on in front of the cameras, but still have a useful continuity file if I ever needed this stuff as evidence.

did you notice the quality issues when encoding to a video stream? or when decoding to still images? I need to post-process a buttload of security footage, and yank 1 frame per second, to then later re-encode that back as a video file for quick scanning and archival processes. I can live with the slower encode times via CPU, but the decode times for each file is pathetic, and that's a problem. I can run up to 4 in parallel right now but that doesn't help the fact it's still days to extract frames from all the footage.

well that's a bummer.

Got some pretty intensive video decode/encode processes I want to do here. thinking about installing this AMD FirePro graphics card, with the hopes of passing it through to a VM to run these FFMPEG tasks quicker.

I'm happily running along on my pair of E5-1230 processor based machines, and 866 plots spread across them. you're reminding me I need to take a backup of the machines soon.

https://chiaforksblockchain.com/ that'll help. also chiadoge is now cryptodoge, with a hard fork of the blockchain.

hang on, I'll have something for you...

looks like shamrock has shit the bed.

and it looks like the 3rd one works. trying now to get everything into their final positions.