Can't create docker...

[jbuszkie]

I'm playing with my unraid and trying to test out a sage install on it.
I have the V8 docker and when I hit create it seems to do nothing. I just have the defaults and didn't modify anything.
Do I need to do something like select a template?
I'm running Unraid 6.1.6.. Do I need to upgrade?

Thanks,

Jim

[Fuzzy]

I'm guessing upgrading wouldn't hurt. I don't think i ever ran unRAID that old (relatively new to it - just since last July I think). that said, it SHOULD work. What it should do when you hit create is pop up a window where it goes through downloading the various images that make up the container, and then launch the container. If you aren't even seeing that happen, I'm thinking it may be an issue with either the unRAID, or your browser. Have you installed any other dockers? do they install properly?

[jbuszkie]

No.. This is my first attempt! I'll try upgrading tonight when I get home.

I've been using unRaid since 2008! lol

[stuckless]

You will need to be on a newer version of unRAID for this to work. If it's an older version of unRAID then what you are seeing (or not seeing) is "normal".

[jbuszkie]

Damn you unraid for being able to update remotely!!!

I just updated Unraid from work and it's pulling sage now! Dammit!! I'm getting nothing done at work today!! :-D

[wayner]

Hopefully you are on a VPN as apparently you should not expose the unRAID UI to the internet.

[jbuszkie]

Yes.. Of course on VPN!

[jbuszkie]

Well... Most things look ok!! It was pretty painless to get the basic Sage
up and running! Now I have to figure out how/when I want to switch over!

Many thanks to Sean and Fuzzy and anyone else that contributed to make this painless!

Right now I see 12 (instead of 6) network encoders. Is that because my current windows sage system is "broadcasting" the 6 prime tuners and it also sees the same 6 that are broadcasted from fuzzy's OpenDCT??

[jbuszkie]

Fuzzy,

You mentioned your specs on the unraid forum... have you had any issues running a parity check and recording max streams? I know when I run a parity check on my current dog, I really can't do much else on it.

I'm worried that I won't be able to handle 6 streams while doing a parity check.. I can't have my shows not record

Jim

[stuckless]

I record to a SSD drive and then move to the array later (ie, my recording share is setup using a cache drive).

Typically running a parity check is not something you'll be doing all the time... I have mine setup to run monthly, and it takes about 11 hours to complete.

[jbuszkie]

Originally Posted by stuckless

I record to a SSD drive and then move to the array later (ie, my recording share is setup using a cache drive).

Typically running a parity check is not something you'll be doing all the time... I have mine setup to run monthly, and it takes about 11 hours to complete.

Mine take about 21 hours to run a parity check. I hope that will get a lot better when I upgrade everything!

How do you move from the SDD cache to the array? Do you use the regular builtin mover script? Or some sort of rsynch

[Fuzzy]

Originally Posted by jbuszkie

Mine take about 21 hours to run a parity check. I hope that will get a lot better when I upgrade everything!

How do you move from the SDD cache to the array? Do you use the regular builtin mover script? Or some sort of rsynch

I use the regular mover script, but have it scheduled to run hourly instead of daily. (FWIW, the mover script IS essentially using rsync). Recordings go to the SSD, and any that are complete will get moved to the array every hour. If a file is being recorded, comskipped, or watched, mover will skip it. Sage never sees it happen, which is the real beauty of the cache mechanic.

Your Parity check taking a long time is probably limited by your drive speeds - it will run at the read speed of the slowest drive in the array, so it is probably wise to have an idea of the individual drive performances (there's a great script on the unraid forums for measuring and displaying all the drive speeds). Gives you and idea of where the weak links are.

[Tiki]

Does unraid let you break up the parity check task?
I currently use flexraid / tRaid, and it allows you to break it up. I have mine set to parity check something like 5% of the array every night. So, it scans the full array about every 20 days and takes about 45 minutes per day.

By the way, size of drives matters too - I have four 4TB data drives and one 4TB parity drive.

[Fuzzy]

No, it doesn't break up the task.. That said, I've also never seen the parity check affect actual operation in any way. If the server is busy with a bunch of access, the parity check just takes longer.

[stuckless]

Originally Posted by jbuszkie

Mine take about 21 hours to run a parity check. I hope that will get a lot better when I upgrade everything!

How do you move from the SDD cache to the array? Do you use the regular builtin mover script? Or some sort of rsynch

If you set a share as a cache drive, the moving happens autoamagically, usually on a schedule, but it should really transparent.

How long the parity check takes will depend on drive speed and how many drive. I have 9TB array where by Parity drive is 3gb in size.

[jbuszkie]

Originally Posted by stuckless

If you set a share as a cache drive, the moving happens autoamagically, usually on a schedule, but it should really transparent.

How long the parity check takes will depend on drive speed and how many drive. I have 9TB array where by Parity drive is 3gb in size.

I think it's more you set the share to use the cache drive. Because there is a cache only option as well as the no cache option. I just wanted to make sure you weren't using some cache only with some manual mover..

My array is 27TB with 12 drives. (9 data, 1 parity and 2 cache) My parity is 8TB.

So my system slows down quite a bit with all that I/O.. Other's were getting better parity check speeds with newer hardware. I'm hoping for the same.

[jbuszkie]

Originally Posted by Tiki

Does unraid let you break up the parity check task?
I currently use flexraid / tRaid, and it allows you to break it up. I have mine set to parity check something like 5% of the array every night. So, it scans the full array about every 20 days and takes about 45 minutes per day.
By the way, size of drives matters too - I have four 4TB data drives and one 4TB parity drive.

Like fuzzy said.. No breaking it up...
We've been asking for a parity check pause for a long time now and I don't think we'll get it. I have mine do it once a month.. but I have a big system so it does take a while..

Jim

[jbuszkie]

Originally Posted by Fuzzy

[From UNRAID Forum]For reference, I am running on a somewhat overkill*i7-6700, but it allows me to transcode the streams from my HDHR-2US and HDHR-Prime on the fly as they are being recorded (I'm using openDCT to interface Sage to the HDHR's, which adds some features, like the transcoding). *I'm using the*ultrafast720p30fps transcode_profile.

With a recording going, watching the processor load, comskip uses about 15-20% of a core, and the transcode uses about 90% of a core.

With 4 recordings going, overalll my unRAID server is at 65% total cpu load, with realtime transcoding and comskip of all recordings. *This is recording to a cache pool made up of 2 mid-tier M.2 128GB SSD's.

What MB are you using? Why the 2 M.2 cache drive instead of a regular sata drives?

[Fuzzy]

Originally Posted by jbuszkie

What MB are you using? Why the 2 M.2 cache drive instead of a regular sata drives?

M.2, depending on the types, use PCIe x4 instead of SATA. A bit faster that way. The sustain 350+MB/s reads each, and don't compete with the SATA controllers that serve the array drives. The MB (GA-Z170X-UD3) had 2 M.2 slots so figured I'd play with them. Also saves drive bays in the case to actual data drives.

[EnterNoEscape]

Originally Posted by Fuzzy

M.2, depending on the types, use PCIe x4 instead of SATA. A bit faster that way. The sustain 350+MB/s reads each, and don't compete with the SATA controllers that serve the array drives. The MB (GA-Z170X-UD3) had 2 M.2 slots so figured I'd play with them. Also saves drive bays in the case to actual data drives.

The SATA bus is maxed out by moderately priced SDD's now. The only way to get more performance is PCIe (or M.2 which is just a fancy PCIe slot). I have a bunch of very fast SATA SSD's and I will never buy another one because while the ones I have are fast, any replacement wouldn't be an upgrade because I've already maxed the bus out. If I buy anything new it's going to be NVMe.