JohnO Posted June 5, 2015 Share Posted June 5, 2015 I am booting from vmdk, not plop. Maybe that's why? I boot from plop. No errors for me. John Link to comment
JohnO Posted June 5, 2015 Share Posted June 5, 2015 I've tried both plop and vmdk What USB devices are available to your unRAID Guest? Are you passing through a complete USB bus, or just individual devices? Are they passed-through and dedicated, or are they shared and assigned at run-time? Are there USB 2.0 devices? USB 3.0 devices? John Link to comment
bonienl Posted June 5, 2015 Share Posted June 5, 2015 By the way, bonienl, I believe you indicated that a way to call for the local generation of the diagnostics dump from the command line was coming. If it's ready now, can you enlighten us how to do it? I'd like to add that to the 'Need Help' post, for when the network is down. The script diagnostics is used to collect the information. The webGUI calls this script in the background with one parameter: the name of the zip file. When telnetting into the system (e.g. when the webGUI is non-responsive), the script can be executed without any parameters. This will produce a zip file with the same naming convention as used in the GUI and stores the file under /boot/logs (the folder will be created if not existing) *Note for LT: still need to symlink the diagnostics script so users can simply type the script name without the full path. And while I have you, a few more requests? Could you change the 'array' folder name to 'system', keep array.txt in it, add system.txt with the results of 'lspci', 'lsscsi', and 'free -mt' (maybe others later?), and add network.txt with the results of 'ethtool -i eth0', 'ethtool eth0', and 'ifconfig'. Of course, you may have better ideas! I'll have a look at that, thanks for the suggestions. Link to comment
sureguy Posted June 5, 2015 Share Posted June 5, 2015 I've tried both plop and vmdk What USB devices are available to your unRAID Guest? Are you passing through a complete USB bus, or just individual devices? Are they passed-through and dedicated, or are they shared and assigned at run-time? Are there USB 2.0 devices? USB 3.0 devices? John I'm only passing individual devices through. I'm not sure how to determine if they're dedicated (sorry about that), but I can't assign them to other guests, once they're assigned (so I assume they're dedicated). I've tried USB 2, and USB 3, with both PLOP and vmdk. The motherboard is an Asrock Z77 Extreme 4, and I've tried USB2 and USB3 ports, but the manual is less than helpful when it comes to identifying which ports belong to which controller. No matter what I do, rolling back to 6beta15 stops the problem. At this point I think it's probably a USB driver issue in the kernel, but that's just random speculation. Link to comment
JohnO Posted June 5, 2015 Share Posted June 5, 2015 The motherboard is an Asrock Z77 Extreme 4, and I've tried USB2 and USB3 ports, but the manual is less than helpful when it comes to identifying which ports belong to which controller. I've got an ASRock motherboard as well. Mine is an AMD board. I think you are probably right that something changed at the unRAID Linux level with the drivers. I've got a screen shot below of my unRAID VM configuration, showing the devices in use. I also have a screen shot of the DirectPath system-wide configuration, showing devices that are dedicated to particular VMs. In my case, I have a storage controller dedicated (Passed-through) the the unRAID VM, but the USB I'm using is just assigned at run-time, and not dedicated (Passed-through) to the Unraid VM. That DirectPath screen shot does show that I have other items passed through to other VMs. Good Luck! John Link to comment
enetec Posted June 5, 2015 Share Posted June 5, 2015 To all of you using unRAID in ESXi... are you using "usb inserted trick" to load unRAID from a virtual disk and "making it happy" with licence by USB key? Did you have any issue with user shares using RC3/4 in this way? Link to comment
OopsAllBerrys Posted June 5, 2015 Share Posted June 5, 2015 I am also using plop and just attach the USB device to Unraid; not passing the whole controller. The "error" usually comes up once an hour. So far I haven't experienced any problems. Link to comment
JohnO Posted June 5, 2015 Share Posted June 5, 2015 To all of you using unRAID in ESXi... are you using "usb inserted trick" to load unRAID from a virtual disk and "making it happy" with licence by USB key? Did you have any issue with user shares using RC3/4 in this way? I don't know what the "usb inserted trick" is. I used PLOP which re-directs the rest of the boot process to the actual USB memory stick. It all seems happy after that. Link to comment
htpcnewbie Posted June 5, 2015 Share Posted June 5, 2015 I have an asrock z97 extreme 4 motherboard. After moving from beta 12 (or maybe 15), all my USB 3 ports that are attached to an on board usb Intel controller are not recognized. I was passing the entire usb controller which has both usb 2 and 3 ports to my kvm and the USB 3 ports do not work anymore (usb 2 ports on the same controller works). I have confirmed that these usb 3 ports do not work even with stock unraid without plugins (in safe mode). I was initially speculating that my 7 month old motherboard had gone bad... Any other suggestions to debug the whether it is the kernel issue? I am currently on rc4. The motherboard is an Asrock Z77 Extreme 4, and I've tried USB2 and USB3 ports, but the manual is less than helpful when it comes to identifying which ports belong to which controller. I've got an ASRock motherboard as well. Mine is an AMD board. I think you are probably right that something changed at the unRAID Linux level with the drivers. I've got a screen shot below of my unRAID VM configuration, showing the devices in use. I also have a screen shot of the DirectPath system-wide configuration, showing devices that are dedicated to particular VMs. In my case, I have a storage controller dedicated (Passed-through) the the unRAID VM, but the USB I'm using is just assigned at run-time, and not dedicated (Passed-through) to the Unraid VM. That DirectPath screen shot does show that I have other items passed through to other VMs. Good Luck! John Link to comment
reluctantflux Posted June 5, 2015 Share Posted June 5, 2015 Boot from a Live CD for some flavor of Linux you're familiar with and check to see if they ports are working in there. Mine are working fine. I am having issues with hard boots messing up my graphics card so bad, that I have to re-seat it. Just unplugging the power to the machine and discharging static power doesn't fix the issue. Seems like more of a graphics card issue than a Unraid issue, though. That was a fun one to troubleshoot. I'm on RC4 now. Link to comment
RobJ Posted June 5, 2015 Share Posted June 5, 2015 The script diagnostics is used to collect the information. The webGUI calls this script in the background with one parameter: the name of the zip file. When telnetting into the system (e.g. when the webGUI is non-responsive), the script can be executed without any parameters. This will produce a zip file with the same naming convention as used in the GUI and stores the file under /boot/logs (the folder will be created if not existing) *Note for LT: still need to symlink the diagnostics script so users can simply type the script name without the full path. Thank you! Can you provide the command a user would use to set the symlink? I located it at /usr/local/emhttp/plugins/dynamix/scripts/diagnostics, and copied it to my flash for easier typing! When I run it without parameters, it stops with the following, required a Ctl-C to break out of, and nothing was created. I have no Shares, nothing in the config/shares folder. It works fine from the webGui. root@JacoBack:/boot# diagnostics cp: cannot stat ‘/boot/config/shares/*.cfg’: No such file or directory ^C root@JacoBack:/boot# Just ideas, would this work? (no PHP experience, so borrowed heavily from you!) Seems to work fine on my system. #!/usr/bin/php <?PHP /* Copyright 2015, Bergware International. * Copyright 2015, Lime Technology * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License version 2, * as published by the Free Software Foundation. * * The above copyright notice and this permission notice shall be included in * all copies or substantial portions of the Software. */ ?> <? if ($argv[1]) { $zip = $argv[1]; } else { exec("mkdir -p /boot/logs"); $var = parse_ini_file("/var/local/emhttp/var.ini"); $zip = "/boot/logs/".str_replace(' ','_',strtolower($var['NAME']))."-diagnostics-".date('Ymd-Hi').".zip"; } $diag = basename($zip, '.zip'); exec("mkdir -p /$diag/array /$diag/system /$diag/config /$diag/log /$diag/shares /$diag/smart"); file_put_contents("/$diag/array/array.txt", str_replace("\n","\r\n",print_r(parse_ini_file('/var/local/emhttp/disks.ini',true),true))); exec("lsscsi|todos >/$diag/system/lsscsi.txt"); exec("lspci|todos >/$diag/system/lspci.txt"); exec("free -mt|todos >/$diag/system/free-mt.txt"); exec("ethtool -i eth0|todos >/$diag/system/ethtool-i_eth0.txt"); exec("ethtool eth0|todos >/$diag/system/ethtool_eth0.txt"); exec("ifconfig|todos >/$diag/system/ifconfig.txt"); exec("cp /boot/config/*.{cfg,conf,dat} /boot/config/go /$diag/config"); if (file_exists("/boot/config/shares/*.cfg")) exec("cp /boot/config/shares/*.cfg /$diag/shares"); $shares = parse_ini_file("/var/local/emhttp/shares.ini",true); foreach ($shares as $share) { $file = "/$diag/shares/${share['name']}.cfg"; if (!file_exists($file)) file_put_contents($file,"# This share has default settings.\n"); } foreach (glob("/var/log/syslog*") as $file) { exec("todos <$file >/$diag/log/".basename($file).".txt"); } exec("ls -l /dev/disk/by-id/[au]*|awk '$0!~/-part/{split($11,a,\"/\");print a[3],substr($9,21)}'", $devices); foreach ($devices as $device) { $disk = explode(' ',$device); exec("smartctl -a /dev/${disk[0]}|todos >/$diag/smart/${disk[1]}.txt"); } exec("zip -qmr $zip /$diag"); ?> Link to comment
limetech Posted June 5, 2015 Author Share Posted June 5, 2015 *Note for LT: still need to symlink the diagnostics script so users can simply type the script name without the full path. That's in next release. Link to comment
enetec Posted June 6, 2015 Share Posted June 6, 2015 To all of you using unRAID in ESXi... are you using "usb inserted trick" to load unRAID from a virtual disk and "making it happy" with licence by USB key? Did you have any issue with user shares using RC3/4 in this way? I don't know what the "usb inserted trick" is. I used PLOP which re-directs the rest of the boot process to the actual USB memory stick. It all seems happy after that. Thanx a lot! Tomorrow I'll test it in Virtualbox too Link to comment
pengrus Posted June 6, 2015 Share Posted June 6, 2015 Quick question, is it normal for the plugins page not to show an update for the Server OS if I'm still back on 10a? Thx! Link to comment
RobJ Posted June 6, 2015 Share Posted June 6, 2015 Quick question, is it normal for the plugins page not to show an update for the Server OS if I'm still back on 10a? That was added in -beta8, but I believe there were a few fixes along the way. Version -beta10a is old, I'd download -rc4 and upgrade manually one more time. Link to comment
BRiT Posted June 6, 2015 Share Posted June 6, 2015 Quick question, is it normal for the plugins page not to show an update for the Server OS if I'm still back on 10a? Thx! Somewhere along the way they switched the GITHUB location, so if you're running a version looking at the old location it will never find an update so you will need to manually update. Link to comment
pengrus Posted June 6, 2015 Share Posted June 6, 2015 Quick question, is it normal for the plugins page not to show an update for the Server OS if I'm still back on 10a? Thx! Somewhere along the way they switched the GITHUB location, so if you're running a version looking at the old location it will never find an update so you will need to manually update. Argh! Tricky Tom & co. Roger that, had hoped to avoid a manual update, but will soldier on. Thanks guys! Link to comment
RobJ Posted June 6, 2015 Share Posted June 6, 2015 I've noticed on RC4 the page update frequency feature set to real time does not dynamically update my drives read/writes until i refresh the webpage. This is from multiple pcs/browsers. Any ideas? Read/Write counters are updated every minute, this hasn't changed. I have performed a test by transferring 250gb to disk3 in my array and waited 15 minutes with disk 3 still showing as spun down. After a manual refresh on my chrome tab I then see disk 3 spun up. Of note before upgrading to RC4 I formatted my USB key, installed Unraid and copied my config folder over. Is there something legacy within my config folder possibly causing the issue? I tested this and it is working as expected in RC4. Counters and disk status are updated every minute. Can you clarify what is actually happening now, and the interaction between poll_attributes and Page update frequency? When you say 'updated every minute', which counter is that? I had previously changed poll_attributes to 180. I decided to see what the new default was (test if it was 60), so I blanked it and hit Apply, and it changed to 1800 (30 minutes). Would it not be better to be about a minute or so, except during a parity check/build, or am I misunderstanding its use? Incidentally, when I went back to the Main screen (I can't say for sure the setting change is the cause), there were temps showing for all but one drive, yet all showed grey balls, spun down. I tested all of them with "hdparm -C", and all but one were "active/idle", the true spun down drive was "standby". There must be conditions where the displayed spin status becomes out of sync with the true spin status. After several hours, long after they should have spun down, they were still showing temps that were drifting up an down. I brought it back in sync by clicking the Spin Up button. Link to comment
Darqfallen Posted June 6, 2015 Share Posted June 6, 2015 Any reason my dockers cannot reach a DNS server but my unraid can? Link to comment
reggierat Posted June 6, 2015 Share Posted June 6, 2015 If you put your server to sleep the status will definitely be out of sync when it wakes preventing the server from sleeping again until a disk is spun up. Link to comment
bonienl Posted June 6, 2015 Share Posted June 6, 2015 Thank you! Can you provide the command a user would use to set the symlink? No need, it will be corrected in the next version. Meanwhile type the full path: /usr/local/emhttp/plugins/dynamix/scripts/diagnostics When I run it without parameters, it stops with the following, required a Ctl-C to break out of, and nothing was created. I have no Shares, nothing in the config/shares folder. It works fine from the webGui. Both CLI and GUI is working for me. Likely it has to do with the missing Shares folder. In the new version more checks are done and should correct your issue. Just ideas, would this work? (no PHP experience, so borrowed heavily from you!) Seems to work fine on my system. Yes, with a few adaptations it will work. I have included the retrieval of the extra information in the next version. Thanks for your suggestions. Link to comment
bonienl Posted June 6, 2015 Share Posted June 6, 2015 Can you clarify what is actually happening now, and the interaction between poll_attributes and Page update frequency? When you say 'updated every minute', which counter is that? I had previously changed poll_attributes to 180. I decided to see what the new default was (test if it was 60), so I blanked it and hit Apply, and it changed to 1800 (30 minutes). Would it not be better to be about a minute or so, except during a parity check/build, or am I misunderstanding its use? Incidentally, when I went back to the Main screen (I can't say for sure the setting change is the cause), there were temps showing for all but one drive, yet all showed grey balls, spun down. I tested all of them with "hdparm -C", and all but one were "active/idle", the true spun down drive was "standby". There must be conditions where the displayed spin status becomes out of sync with the true spin status. After several hours, long after they should have spun down, they were still showing temps that were drifting up an down. I brought it back in sync by clicking the Spin Up button. There are different parts dynamically updated: GUI The updating/refreshing of the GUI happens at the frequency set for Page update frequency, default is real-time. Whenever the GUI is refreshed it will read all stored values of emhttp of that moment. emhttp values The emhttp values are updated in the background by the monitor process or in the foreground when the user changes a page in the GUI. The background updating is done every minute. disk temperatures Disk temperatures are handled separately. smartctl is called at the frequency set for Tunable (poll_attributes), default is 30 minutes. The reason this is done at a much lower frequency is to minimize possible interruption of normal disk operation while reading the disk temperature. Link to comment
bonienl Posted June 6, 2015 Share Posted June 6, 2015 Quick question, is it normal for the plugins page not to show an update for the Server OS if I'm still back on 10a? Thx! Somewhere along the way they switched the GITHUB location, so if you're running a version looking at the old location it will never find an update so you will need to manually update. Correct, this happened when Dynamix was introduced as GUI in v6b12. Link to comment
trurl Posted June 6, 2015 Share Posted June 6, 2015 There are different parts dynamically updated: GUI The updating/refreshing of the GUI happens at the frequency set for Page update frequency, default is real-time. Whenever the GUI is refreshed it will read all stored values of emhttp of that moment. emhttp values The emhttp values are updated in the background by the monitor process or in the foreground when the user changes a page in the GUI. The background updating is done every minute. disk temperatures Disk temperatures are handled separately. smartctl is called at the frequency set for Tunable (poll_attributes), default is 30 minutes. The reason this is done at a much lower frequency is to minimize possible interruption of normal disk operation while reading the disk temperature. I love it when something vague and mysterious is clarified. Thanks! Link to comment
Recommended Posts