Few questions after take off.

[pwm]

The errors are normally from a bad cable or a cable that hasn't been properly plugged in. UDMA CRC errors can't be reset. You can acknowledge the error on the dashboard. Then unRAID will warn if the value changes.

 

Check that the cable is properly fitted and if the counter values stays then you are good to go. Else consider getting a new SATA cable.

 

[shteud]

1 hour ago, pwm said:

The errors are normally from a bad cable or a cable that hasn't been properly plugged in. UDMA CRC errors can't be reset. You can acknowledge the error on the dashboard. Then unRAID will warn if the value changes.

 

Check that the cable is properly fitted and if the counter values stays then you are good to go. Else consider getting a new SATA cable.

 

sorry for my bad english. 

I've already replaced SATA cable (twice :)) and perform "extended smart test" in unraid. And it says "complete without errors". 

But the errors still is shown in the table.

but now i get how it's work. 

so it's must be all ok for placing this HDD in the pool (anyway preclear procedure i think show any issues with this disk)

[Squid]

2 hours ago, shteud said:

after i replace it extended smart test show no troubles. but the error attribute doesn't disappear from the table. should i worry about it?

The attribute is a counter.  It does not reset to zero, simply increases whenever that error appears.

[shteud]

pwm  Squid

Thanks !

Edited December 30, 2018 by shteud

[shteud]

Hello all again!

Don't want to start a new topic so just ask here.

A had a quite good ride with my unraid setup, and now i plan to add some more space.

And i have a little dilemma.

Now i have 2x 2TB toshiba 7200 rpm 64MB set for 1 for parity+1 for the data.

And i bought new seagate 2TB 7200 rpm 256 MB cache and it's obviously faster one.

So don't really sure should i replace parity disk or use expand data pool?

Where this extra speed bee more helpful?

Or it's hasn't any matter because of ssd cache?

[trurl]

Use the faster disk to replace parity then use the old parity to expand data.

[shteud]

On 12/30/2018 at 3:33 PM, trurl said:

Use the faster disk to replace parity then use the old parity to expand data.

Thank you!

One more little question.

How can i check, writing resource of ssd in cache array? Only smart? 

Maybe i can somehow interpret data from this row?

[trurl]

Not sure I understand the question, but the numbers in your screenshot will reset on reboot.

[shteud]

47 minutes ago, trurl said:

Not sure I understand the question, but the numbers in your screenshot will reset on reboot.

Sorry 

A little bit tired, think slow

I mean how i can check ssd flash memory resource in unraid? In windows, we have a lot utilities. But have no idea what to use for this purpose in unraid.

[trurl]

You can check SMART Attributes and do short and extended SMART tests of any disk by clicking on it in the webUI to get to its page.

 

If you have something else in mind you will have to spell it out to me or wait for someone else who knows what you mean.

[shteud]

 

trurl 

Thanks. I think i get that i need.

2 ALL

I have one more question dedicated to hardware.

A few days ago i download on NAS with unraid, 3 big torrent files (total size about 70 GB and download speed was about 40MB/s) with transmission client (docker) and in same time i try to watch film from NAS (it was about 30 GB) . And i should say it was impossible. Playback was very laggy and unstable (i looked at unraid dashboard and seems sometimes CPU got 100% load at all cores).  I wondering do this problem occurred via low horsepower under the hood or some kind of software issues? Now i use with unraid Intel Core i3-2130 CPU @ 3.40GHz/ 12GB RAM/120GB ssd cache/[email protected] disk pool. 

I think to try to change cpu on something with 4 cores. How do you think it solve the problem? And if yes, do i need take 4c\4t or 4c\8t CPU? Unraid can effectively utilize hyper-threading? 
Do i need more ram?

Edited January 18, 2019 by shteud

[ken-ji]

Are you downloading to the cache drive or the array drive?

Because downloading to the array can result in I/O starvation - parity is being read and written to, array disk is being written to, (maybe a different) array disk being read from (movie stream) - and if happens, this can cause your disk access to be very slow and make CPU usage look a lot higher - the dashboard silently includes I/O wait time as CPU usage.

 

Some users, like me, download to a cache drive so unRAID will handle disk I/O better.

[shteud]

On 1/21/2019 at 3:10 AM, ken-ji said:

Are you downloading to the cache drive or the array drive?

Because downloading to the array can result in I/O starvation - parity is being read and written to, array disk is being written to, (maybe a different) array disk being read from (movie stream) - and if happens, this can cause your disk access to be very slow and make CPU usage look a lot higher - the dashboard silently includes I/O wait time as CPU usage.

 

Some users, like me, download to a cache drive so unRAID will handle disk I/O better.

Yes, this time it was direct write to HDD.  But read was going from other HDD.  
Now I'm thinking, did CPU upgrade will help in scenarios like this?

[ken-ji]

Please note that there is distinction between a direct write to array disks vs cache disks.

When a write to array disk occurs, Unraid needs to read the parity disk + target array disk, then using that compute the new parity value, and then write to the target array disk and the parity disk. If you try to read from any drive during this time, you'll have disk I/O contention at a 4:1 ratio. It won't be a problem just yet, but you will encounter stuttering and pauses which depends on your hardware and the speed the download is ocurring.

 

As for the CPU, your i3 is more than enough to do the parity computation. I only use a Pentium G4620. Upgrading CPU won't make a difference.

 

You didn't provide your complete specs or the diagnostics file, so you'll only get general tips

If you have a spare HDD (or SSD which is even better), you can add it and assign it as a cache drive.

This speeds up your download as it will only need to write to cache drive like in any normal system. But files on the cache will only get moved to array on a schedule

 

You also didn't consider the state of your network, is Unraid connected via gigabit? 

[shteud]

Thank you ken-ji!

I understand how caching work in unraid. I'm pretty sure i setup transmission to save data on cache first. But i think this time it was direct write to HDD (idk why :() 

So no sense for an upgrade? Only caching can help in a situation like this?

In the quote, you can see my Unraid server specs

Quote

System Overview

unRAID system:unRAID server Basic, version 6.6.6

Motherboard:ASUSTeK COMPUTER INC. - P8H77-M

Processor:Intel® Core™ i3-2130 CPU @ 3.40GHz

HVM:Enabled

IOMMU:Disabled

Cache:L1-Cache = 128 kB (max. capacity 128 kB)

L2-Cache = 512 kB (max. capacity 512 kB)

L3-Cache = 3072 kB (max. capacity 3072 kB)

Memory:12 GB (max. installable capacity 32 GB)

eth0: 1000Mb/s, full duplex, mtu 1500

P + Q algorithm:9800 MB/s + 14191 MB/s

 

SSD CACHE 

2x120 GiB sata3 ssd

 

HDD

parity ST2000DM008-2FR102_WFL0P7LR - 2 TB (sdf)

data 2xTOSHIBA_DT01ACA200_53F8XUGGS - 2 TB (sde)

 

A home network is 1GB/s on SMB class switch + pfsense router. Internet is 1GB/s too. 

 

1

 

[trurl]

Diagnostics will give us a better idea how you have your user shares cached.

[shteud]

8 hours ago, trurl said:

Diagnostics will give us a better idea how you have your user shares cached.

Ok np. Here is a file.

nas-server-diagnostics-20190125-2123.zip

[trurl]

Your system share is showing it has files on disk1 so it could be your docker image is there instead of on cache. If you enabled docker before installing a cache disk this is likely.

 

You can see which disks a share is using by going to the Shares page and clicking Compute... for the share.

 

Go to Settings and disable the docker and VM services, run mover, then check to see if system share still has any files on disk1.