Dynamix - Web GUI

[bonienl]

So, disabled is my friend 

 

Your stats show less than 1% CPU utilization in active/passive mode, why your choice for disabled mode ?

[bonienl]

my suggestion is to have passive as default on next release.

so half of the ppl tat use this plugin wont cry when they r slow during parity check

 

People should read first the OT before complaining, it is explained what these modes do and it is a small step to check and change as needed...

[garion]

So, disabled is my friend 

 

Your stats show less than 1% CPU utilization in active/passive mode, why your choice for disabled mode ?

 

Hello,

 

I´m not a linux expert but I notice the %cpu increase in the stats plugin. It shows min 20% and peaks about 40% in User %CPU. Then I unraid first (Top info) and telnet later and begin tests. If I´m not wrong its not a 1% (I don't spend two hours testing for that  ). In top we have the 1-5-15 min average CPU use, what I wrote in bold in my previous post and as I understand (and read) Active values are not really ok. I have a dual core so 100% CPU load its 2.00, 1.00 its 50% CPU load and as you can see TOP values are in line with it was shown by Stats plugin 0.70 average in 5 minutes  its a 35% CPU load in my dual core.

 

But I might be wrong. 

 

Other is that I notice a not usual behavior or my parity drive. It keeps spinning. If I push "spin down" button parity spin down but return up in about 20-30 minutes (testing now).

 

I am investigating this as well, because I have never seen or heard about this before (SF uses the same mechanism).

 

The disk health plugin does a background status check every 10 seconds and polls all available disks sequentially, there isn't really a difference between first and last disk (at least there shouldn't be).

 

The background health check is only active when the health page is open, moving to another page view will stop the background health checking. Any disks which are spun down are not health checked unless the option "Enable background polling for spun-down disks" is set to yes.

 

Now a couple of questions:

 

1. The interval after which the parity disk goes "on" is 20-30 minutes ?

2. You observe this while looking at the disk health page (which is opened for the duration you mentioned) ?

3. What setting is "Enable background polling for spun-down disks" ?

4. What is the spin down time of your parity disk ?

5. What brand/model is your parity disk, is it different from your data disks ?

6. Does the parity disk connect directly to the MB or do you use a separate controller card ?

 

 

Since last night I delete disk health parity behavior its correct, no spin-up of any drive itself. I´ll try do more test tonight.

 

To your questions:

 

 

1. The interval after which the parity disk goes "on" is 20-30 minutes ?

I think so, sure more than 10 minutes

 

2. You observe this while looking at the disk health page (which is opened for the duration you mentioned) ?

No, main page or stats page.

 

3. What setting is "Enable background polling for spun-down disks" ?

off

 

4. What is the spin down time of your parity disk ?

same than all array. 30 minutes.

 

5. What brand/model is your parity disk, is it different from your data disks ?

I attach a picture 1 seagate parity, 5 seagate data, 2 WD data, 1 seagate cache

 

6. Does the parity disk connect directly to the MB or do you use a separate controller card ?

MB

 

And this maybe its important: Even with "Enable background polling for spun-down disks" disabled in Health tab if I click in seagate drives that drive spin-up. If a click in a WD drive, that drive show information but stay spin-down.

 

If you want I do some test please ask for it 

 

best regards

[hemmiandra]

I noticed the same thing since installing dynamix. Not one of my disks spin down. I've checked when the fileserver is in minimal or no use at all late at night/really early in the morning and all the disks are spun up, including the parity drive. I have the spin down time set at 1hour, all the drives are connected through sata multipliers and all the drives are 3TB seagates, same model.

 

Is there any way of monitoring spin down/up activity? SMART values perhaps?

[bonienl]

I have one Seagate drive in my system, this one doesn't spin down while all of my WD drives do spin down.

 

This happens because the page refresh mechanism causes emhttp to execute the 'smartctl' program in the background and this prevents the ST drive from spinning down. Once it is spun down, then emhttp won't query the drive any longer.

 

Two options:

 

1. Press the button "spin down" in the webGui

 

2. Disable the page refresh and use manual mode (like the stock GUI)

 

-----------

 

Ideally I like to see a change in the behavior of emhttp. The way it operates now makes it hard to introduce interactivity.

 

[hallifallen]

Hey, I installed this last night and noticed something weird today.. I installed a bunch of the plugins, and when I was checking the User Names in Active Streams I can see a bunch of WAN IP's in the list. I don't know what to make of this.. Are a bunch of people from around the world trying to access my server?

The server is connected 24/7 to a vpn with the OpenVPN plugin, if that makes any difference. Also I have utserver installed for torrenting.

 

**EDIT**: okay, I figured it out. I reinstalled the active streams so the user names table would be reset. Then it only showed my local ip's who were accessing the server. Then as soon as I started a torrent on the utserver the table got flooded with peer IP's from that torrent. No hazard in that, but a weird bug.

[SlrG]

@bonienl:

Great work! 

 

But I have some observations:

I'm often using my netbook with 1024x600 resolution to check my unraid status. I would like the fixed width value to be adjustable by the user. Do you think this is possible? At the moment I have to zoom out to see everything without scrolling sideways, but it gets painfully small that way. 

 

The spundown disks have a lighter shade of green. But once all disks are spundown it is slightly hard to tell which green is showing. The old blinking icon was better, IMHO.

 

Also the double arrow to manually spindown disks seems to be missing. It is mentioned in the popup legend, but I don't see it for the disks.

 

After choosing to stop the array on Array Operation the screen goes back to Array Devices. I think it would be better to stay there, so the user might choose to reboot, powerdown or whatever without having to click on Array Operation again.

 

In front of the parity check duration there is a upside down A symbol. I guess that is due to a wrong font mapping.

 

The Disk Usage shown for the flash drive is still wrong (same situation as with the stock webui and SF) if there is a smaller first partition on the flash. I have an 16gb usb stick with an 1 gb partition which is filled with 527 MB data in use for unraid. (Imaging 1GB for backup purposes is much faster than imaging 16GB).

 

Having the plugins only as txz files on github makes it hard to contribute changes. It would be better if the source was directly available. For instance the s3 sleep plugin. On SF I had it modified in a way, that the user can choose to ignore disks that are not array disks from activity detection. This was necessary, as the normal script checks all disks. My plex drive ssd installed outside the array stopped the s3 sleep from working as it was always reporting an active state. If you are interested I will try to port my changes to your dynamix s3sleep script.

[garion]

Bonienl,

 

S3 works like a charm¡¡¡

 

I only test a few times with the sleep button but server go to S3 state fine, then XBMC with WOL do the magic 

 

It could save me 400 to 500 KW/yr

 

------

 

Problem with the GUI in iPad (missing backgrounds) disappear today. I deleted the cache yesterday but... I know, you told¡¡¡ Now only the width issue is there but its a minor problem.

 

 

[bonienl]

Then as soon as I started a torrent on the utserver the table got flooded with peer IP's from that torrent. No hazard in that, but a weird bug.

 

The current implementation isn't selective enough, it takes all established sessions to create the users list, including non-relevant ones (like utorrent). Next version will have this corrected. Thanks for testing.

 

[bonienl]

I'm often using my netbook with 1024x600 resolution to check my unraid status. I would like the fixed width value to be adjustable by the user. Do you think this is possible? At the moment I have to zoom out to see everything without scrolling sideways, but it gets painfully small that way. 

 

To solve the issue with the stats tables I had to move away from a min-max width setting. It used to be between 960 to 1240, but now is fixed to 1240. Will do some tinkering to see what can be done.

 

The spundown disks have a lighter shade of green. But once all disks are spundown it is slightly hard to tell which green is showing. The old blinking icon was better, IMHO.

 

The blinking icons made me nervous 

 

The original .gif icons are replaced here with .png versions. These have higher image quality and support transparent background, but don't support animation.

 

Personally I don't have a problem to distinguish on and off modes, but I take your comment in consideration - no blinking however.

 

Also the double arrow to manually spindown disks seems to be missing. It is mentioned in the popup legend, but I don't see it for the disks.

 

You have to click now on the ball icon to spin up/down the disk. Move your cursor over the icon and you will see.

 

After choosing to stop the array on Array Operation the screen goes back to Array Devices. I think it would be better to stay there, so the user might choose to reboot, powerdown or whatever without having to click on Array Operation again.

 

Cookies are used to remember the previous selected tab. You need to have cookies enabled for your browser or at least have them accepted for your local site.

 

In front of the parity check duration there is a upside down A symbol. I guess that is due to a wrong font mapping.

 

No that is a "forall" mathematical symbol, meaning the statement is true 

 

The Disk Usage shown for the flash drive is still wrong (same situation as with the stock webui and SF) if there is a smaller first partition on the flash. I have an 16gb usb stick with an 1 gb partition which is filled with 527 MB data in use for unraid. (Imaging 1GB for backup purposes is much faster than imaging 16GB).

 

All implementations read the same "size" value of the flash drive, to get it corrected it needs to be changed by Limetech.

 

Having the plugins only as txz files on github makes it hard to contribute changes. It would be better if the source was directly available. For instance the s3 sleep plugin. On SF I had it modified in a way, that the user can choose to ignore disks that are not array disks from activity detection. This was necessary, as the normal script checks all disks. My plex drive ssd installed outside the array stopped the s3 sleep from working as it was always reporting an active state. If you are interested I will try to port my changes to your dynamix s3sleep script.

 

Yeah I can add the source of all plugins for those who want to make changes or give code contributions (you need now to unpack the txz files to view the source).

 

[bonienl]

There is a (known) bug in unRAID where newly added drives won't spin down automatically.

 

The workaround is to do the following:

 

main -> disk X -> Device settings -> click on "Apply"

 

The above will correctly write all the settings for disk X (this applies to parity and cache as well). The setting for "Spin down delay" isn't written when the drive is added the first time to the array (this is the bug).

 

So people observing that one or more drives are not spinning down, could try the above workaround first.

 

 

[bonienl]

Created a web portal on github to support the project, have a look at my dynamix page.

[peter_sm]

Hi, is this possible to add..

 

1: add System Stats for tun+ and tap+ interfaces

2. Disk Health for Unassigned Devices.

3. s3 plugin: Ignore Unassigned Devices

4. s3 plugin: Function to disable sleep when a telnet session is established

 

Remote = never sleep, even with no session established

Local: sleep when telnet in

 

 

//Peter

 

[peter_sm]

With the new sensors, i see this, but nothing in the web UI

 

sensors
coretemp-isa-0000
Adapter: ISA adapter
Physical id 0:  +35.0 C  (high = +85.0 C, crit = +105.0 C)
Core 0:         +25.0 C  (high = +85.0 C, crit = +105.0 C)
Core 1:         +34.0 C  (high = +85.0 C, crit = +105.0 C)
Core 2:         +23.0 C  (high = +85.0 C, crit = +105.0 C)
Core 3:         +31.0 C  (high = +85.0 C, crit = +105.0 C)

nct6776-isa-0290
Adapter: ISA adapter
in0:           +0.90 V  (min =  +0.00 V, max =  +1.74 V)
in1:           +1.85 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in2:           +3.39 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in3:           +3.38 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in4:           +0.01 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in5:           +1.68 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in7:           +3.44 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
in8:           +3.26 V  (min =  +0.00 V, max =  +0.00 V)  ALARM
fan1:         1826 RPM  (min =    0 RPM)  ALARM
fan2:         2093 RPM  (min =    0 RPM)  ALARM
fan3:            0 RPM  (min =    0 RPM)  ALARM
fan4:            0 RPM  (min =    0 RPM)  ALARM
fan5:            0 RPM  (min =    0 RPM)  ALARM
SYSTIN:        +26.0 C  (high =  +0.0 C, hyst =  +0.0 C)  ALARM  sensor = thermistor
CPUTIN:        +25.0 C  (high = +80.0 C, hyst = +75.0 C)  sensor = thermistor
AUXTIN:        +47.0 C  (high = +80.0 C, hyst = +75.0 C)  sensor = thermistor
PECI Agent 0:  +34.0 C
cpu0_vid:     +0.000 V
intrusion0:   ALARM
intrusion1:   ALARM

[dalben]

Is there a web based preclear wrapper with this webgui ?

[tr0910]

Is there a web based preclear wrapper with this webgui ?

 

Joe L (the author) has never taken this on even with his unmenu.  UnRaid needs an integrated pre-clear to appeal to the drobo crowd. I wonder what has kept this from happening?

 

Sent from my mobile

 

 

[mikejp]

Is there a web based preclear wrapper with this webgui ?

I would guess there won't be... not after this

 

Although I'm sure you think you've made me happy, you have not. 

Please write your own preclear script and support it.  I do not support you using mine....

 

Please remove my code from your plugin.  You do not have my permission to use it.    Write your own d*** script.

 

Joe L.

 

when it was tried with Simple Features 

[bonienl]

Guys lets not start this story again... 

 

Maybe something can be worked out for the future but not for the moment.

[bonienl]

With the new sensors, i see this, but nothing in the web UI

 

sensors
SYSTIN:        +26.0 C  (high =  +0.0 C, hyst =  +0.0 C)  ALARM  sensor = thermistor
CPUTIN:        +25.0 C  (high = +80.0 C, hyst = +75.0 C)  sensor = thermistor
AUXTIN:        +47.0 C  (high = +80.0 C, hyst = +75.0 C)  sensor = thermistor

 

If SYSTIN == motherboard and CPUTIN == processor, it can be added to the detection logic of the temp plugin, but I need a confirmation.

 

Another approach is that you update the sensors.conf file on your system, likely need to search to SimpleFeatures topic to find directions, I don't believe a separate topic exists, but I wasn't here for quite some time ...

[garion]

back to topic  today I reinstall Stats to be sure that plugin wasn't responsable of the always spin-up parity and it doesn´t  . All drives works as they should be. Spin-up/down same way than stock GUI. Disk health guilty¡¡¡ Seagate and their crappy firms more guilty¡¡¡ (I´ll do another test to be 100% sure)

 

Speaking again about active/disable mode  a picture better than one thousand words.

 

To understand the graph:

 

 

From 12:30 to 15:10 "Disable" mode on. Unraid was "streaming" to Raspberry two animation films, average 6 Mbps nitrate, only 1 connection to the server.

From 15:10 to 16:05 Unraid was in a no task mode. All disk spun-down.

From 16:05 to 16:25 In XBMC Rpi I did first a clean video database then a update library. Unraid spin up 4 drives.

From 16:25 to 16:45 I turn on "Active" mode and stream again an animation film, same 6 avg bitrate to Rpi. Same task than 12:30 but now USR %CPU is lot higher. Webgui webpage is open.

From 16:45 to 17:00 Webgui closed.

17:00 Open webgui - go to stats - get picture

 

So:

Active mode has about 10% CPU use in a Intel Dual core G1610. Yesterday test shown about 20% but maybe was due to Disk health plugin, not installed today.

There is no difference if your are in main page or stats tab.

Active mode only increase CPU if webgui is open.

 

I like this GUI

[peter_sm]

 

If SYSTIN == motherboard and CPUTIN == processor, it can be added to the detection logic of the temp plugin, but I need a confirmation.

 

Another approach is that you update the sensors.conf file on your system, likely need to search to SimpleFeatures topic to find directions, I don't believe a separate topic exists, but I wasn't here for quite some time ...

 

It is solved now with the sensors.conf

 

//Peter

[Frank1940]

I installed Dynamix 2.0.2 on my main server today.  I did have an issue you might want to look into.  I loaded seven of the plugin packages on to the flash drive and booted the server.  Everything installed EXCEPT the email package.  I reboot the server and the same condition occurred. I looked in the /boot/config/plugins/dynamix folder and the .cfg file was there.  I copied over the email .plg file again and rebooted again.  It still failed to install.  Next, I copied over the dynamix.email.notify-2.0.1-i486-1.txz file from my testbed server and and rebooted the server.  The email package was now installed. 

 

The only thing I can figure that happened is that the dynamix.email.notify-2.0.1-i486-1.txz failed to download.  What makes this suspicious is that other six .txz files were downloaded.  You might want to check this out...

 

The temperatures (on the main server) are missing so I will have to resolve this issue...

 

I do want to say that the passive mode for the display seems to work perfectly. 

[bonienl]

The only thing I can figure that happened is that the dynamix.email.notify-2.0.1-i486-1.txz failed to download.  What makes this suspicious is that other six .txz files were downloaded.  You might want to check this out...

 

Email was (silently) updated to version 2.0.2, you need to get the latest PLG. Indeed version 2.0.1 txz is not present anymore on github.

 

I am working on a plugin update feature for a next version, this would allow you to see what has been installed and the latest version available, with the possibility to upgrade (kind of plugin manager).

[bonienl]

 

If SYSTIN == motherboard and CPUTIN == processor, it can be added to the detection logic of the temp plugin, but I need a confirmation.

 

Another approach is that you update the sensors.conf file on your system, likely need to search to SimpleFeatures topic to find directions, I don't believe a separate topic exists, but I wasn't here for quite some time ...

 

It is solved now with the sensors.conf

 

//Peter

 

Good to hear.

 

Ps1. Adding an ambient  sensor, I put as a feature request. Thanks for testing.

 

Ps2. You may want to share your solution, it might help others with similar problem.

 

[trurl]

Here is how I got my sensors working:

 

Run sensors-detect. Not sure where sensors-detect came from, is this part of the dynamix plugins, or is it built in to unRAID?

Since sensors-detect is a perl script, I had to install perl from unMenu first to get this step to work.

 

Got sensors.conf for my motherboard from lm-sensors.org and put it in /boot/custom/etc/sensors.conf

 

Ran the following and also added to go

 

#modprobe detected sensors
modprobe it87
modprobe coretemp

#copy the sensor.conf file for use
cp /boot/custom/etc/sensors.conf /etc/sensors.d