Neo_x

must admit - stubborn me did not want to believe it, but seems like temperature was indeed the cause. just the stubborn, errr i mean loyal old server, wanted me to remove it from the server cabinet and put it right next to my bed before it started to complete a parity sync (combined with a massive 75 cm fan for the warmer times of day. i believe there is other possible issues afoot - possibly some of the controllers, as i had 4 drives drop out of the array yesterday during fan troubleshooting, which required me to run xfs_repair before i could mount them again. will monitor it closely i am getting a fan controller in the next week to see if i can manage the temperature a bit better as well. Thx for the help team!

thx guys. i will put a temporary measure on to see if i can drop the temperature. currently air is being sucked over the drives and pushed through cpu and then out, so shouldn;t be an issue. might be that the fans is a bit low CFM, but will try to manage/upgrade witha controller.

Hi team any ideas as to why a parity sync will all of a sudden drop 50% in speed? i am trying to complete a parity rebuild to a new 8TB drive (tested with preclear before the time), and for the first +-13 hours it was performing admirably (roughly 35MB/s). since this afternoon however, it dropped to 10MB/s I dont have anything writing to or reading from the array, so it is really strange to see this : I have disabled anything extra (cahcdirs , and now for troubleshooting even the docker and VM services). No details in syslog as to a caused. can anybody maybe give an idea? not willing to wait 7 plus days for a parity sync to finish at current speeds , but also dont want to give up on 25% / 24 hours alread completed... TIA neo_X storage-diagnostics-20180826-2322.zip storage-diagnostics-20180826-2322.zip

you had me saying "DUHHHHH" out loud in the office. (although i always base it on a percentage). will balance/move some data around tonight to test, but i believe this is definitely the answer! thank you guys!

didn't touch split level for a long time (data equally distributed across the other 18 discs). believe it is level 2? sample directory under a share : share name : Movies Folder 1 : " A" / "B" / "C" etc Folder 2 : "Movie name (year)" currently all the drives contains "Folder 1", with "folder 2" being unique on all the drives.

Hi unraid team i'm having a very very odd problem, where my user share just refuses to utilize one of my drives. - Setup is very stable(8 years stable), and is working fine, even with a drive i added as little as a month ago(Disk#18 in screenshot) (pre-clear and added). the drive giving issues now(Disk#19) was previously a cache drive, but pre cleared before adding to the array. Unraid version 6.4.1 *edit - upgrade to 6.5.2, same issue persists* some of the things i have tried : Check on Global share settings (with array stopped). Drive is enabled (nothing excluded) check on Share settings for the specific shares - all drives is selected. Copy data using MC directlty between drives - opening the user share shows that the files is visible (which makes me believe the drive is part of the share?) Thus the conundrum , why is "Most Free" allocation method not utilizing the drive? (Disk 19 in below screenshot) TIA! Neo_X storage-diagnostics-20180604-2108.zip (version 6.4.1) storage-diagnostics-20180604-2158.zip (version 6.5.2)

Hi Binhex reverse question lol. which torrent providers are you running on MG?

Hi guys i recently upgraded my V5 to V6 ( yes i know - one of the late bloomers ) once booted, everything seems to be really stable - i even went ahead with adding more drives and a secondary parity. my issue comes in on my power management. due to high electricity costs, i usually only power on the server when i want to view some of the linux cd's i have loaded on there on my TV. The issue i picked up recently is that i need to wait a long time for the server to come available on the network after issuing the WakeOn Lan packet ( close to 5 and a half minutes) I connected a monitor today to the server to see whats going on : SAS cards intializes in about 1.5 minutes. then the BZIMAGE and BZROOT starts unpacking. this takes close to three minutes if not longer, after which the boot resumes and completes loading the docker and VM. i seem to remember, especially in V5,especially the later versions, that BZROOT and BZIMAGE unpacked almost instantly ( <10 seconds), which is why its odd to go back to the slow unpack speed. i already checked the USB on my windows machine for errors - nothing found. a quick speed test on the drive reports about 15MB/s - which should be more than fast enough to read through the <100MB BZROOT file. Can anyone confirm if they are experiencing the same, and maybe recommend how i could troubleshoot this? Thx Neo_X

i was hoping that this will be possible (always a big issue to determine where a parity error is coming from (and to go through mutiple smart reports manually to determine/test possible cause is difficult. what other alternative is there to avoid "bit-rot" (if i am using the correct term) (on that same topic - will Reiser or XFS be recommended as a more stable file system?) If you are concerned about bit-rot then use btrfs and ECC RAM. If not so much, use xfs. These days I think it's wise to avoid ReiserFS. (Meaning: I don't think anyone needs to rush and convert all their existing reiserfs volumes - I still have a large number of them - but for all new storage devices, btrfs or xfs is the way to go.) Thank you sir! yes makes sense (backups first, then file integrity , but also then improving the hardware and file system seems to be a way to battle it). i recently encountered a few parity errors (after having 0 for the past year), but i think its probably due to a power failure while writing to array (mover was running). will investigate with a few repeat checks to make sure hardware is still co-operating.

i was hoping that this will be possible (always a big issue to determine where a parity error is coming from (and to go through mutiple smart reports manually to determine/test possible cause is difficult. what other alternative is there to avoid "bit-rot" (if i am using the correct term) (on that same topic - will Reiser or XFS be recommended as a more stable file system?)

Facepalm, totally missed that. That'll teach me to read fast before coffee. let me guess . you read past it at........ warpspeed?

finally a DONATE button! asked for it years ago in the SAB thread - : - ) -> sent you a few beers doesnt seem like the version number has been updated. wondering if there has been any improvements released after 2.2.0b9?

Nice review above! unfortunately i have had a bad experience, which i am not 100% sure it is related to the plugin but here goes. I helped a friend build a mini server the last week or two, where we picked up a HP gen8 micro server on special, to which we added 3 seagate 8TB archive HDD's. I managed to get the preclear working via the gui after creating an array , and then deleting the config again ( Preclear was complaining about a missing disk.cfg). in any event - we then connected to the gui every few hours - monitoring the progress, and everything went very ok . Preclear finished in about 72 hours which is roughly double the time my 4TB drive took - really ok with that - especially since we performed all three drives at the same time. Problem we encountered - directly after the preclear, the main page showed some strange values for the flash drive. Flash ID was missing ( with unraid requesting us to buy a key again), Flash reported reads of something in the range of 144,000,000 (massive number - i might be missing some zero's ), and writes of about 16,000. restarting the machine confirmed my suspicion - the Flash gave up on us (in hindsight - i should have made a screengrab and captured the syslog) Is there any chance that the plugin performs constant reads from the Flash in any way? i dont think there should be a reason, but just in case. going to return the flash today, and request a replacement key from Limetech - so server should be able to get going soon.

it was 42 wasn't it?

glad you got sorted. this is exactly the users i think Limetech wants to target as well as business installations. My unraid is stable as nothing before - thus having a quick experienced engineer setup a new install wont be a pain at all for many users out there. keep the good times coming!

hi guys i have recently upgraded the s3_sleep plugin to the latest version, as well as unraid from 6.0.1 to 6.1 RC2. problem is, after that upgrade, i have issues all -over. mostly with drives not spinning down, or if they have spun down, s3_sleep still picks them up as running (as per dashboard status) is there any way to try and troubleshoot? I have rolled back to 6.0.1 some captures below : diagnostics with syslog is attached. Please advise what other troubleshooting i can maybe do? thx Neo_x storage-diagnostics-20150811-2326.zip

Attributes (temperature) are only polled every 30 minutes by default. This means that if a disk spins down, it can take up to 30 minutes for the temperature to turn to "*". I have mine set to poll every 5 minutes, for more responsive updates. You can tune this under Disk Settings. Not the issue here. I usually find the 'problem' in the morning when I get the Status Reports e-mails. The report, with a time stamp of 20 minutes after midnight, will say the disk(s) are active. I will then start the GUI (this now being after 7:00AM), the disk temperature(s) are there and the indicator says the disk(s) are in the 'Stand By Mode (spundown)'. So it is not a matter of Attributes polling not having time to 'synced up' with the actual disk state. PLUS, I have set the Polling time to 60 seconds since I don't keep the GUI open unless I am actively using it! There are 2 things I know of that cause the out-of-sync temps to spin status - * using s3 sleep plugin and waking up causes the drives to be spun up without informing emhttp, so the drives show temps but also show spun down * clicking on a drive brings up the drive info, Check Filesystem, and SMART info sections; it will often briefly indicate something about only showing SMART info when drive is spun up, then will apparently spin the drive up(!) and show all of the SMART info; on return to the Main page, drive still appears spun down but the temp is showing * I'm sure there are other causes, be good to identify them Mine too is set to 60 second update. It looks like emhttp assumes it knows all, and trusts all other agents to inform it of spin up or down, which obviously isn't happening. It needs to check for itself the true physical state, much more often, or find ways to be better informed of spin ups. i upgraded to RC2 last night, and are having similar issues. about 5 of my Western digital drives (3TB) shows temperature, but according to dashboard it is spun-down. i usually run s3_sleep to shutdown (not sleep..) the server after detecting that the drives is spun down.. this is however not happening. i have set the polling frequency under disk settings to 60 seconds, and forced a spin-down. it still shows one drive with a temperature, and s3_debug shows the following : any other way to troubleshoot this bug further? i'll roll-back to stable in about 2 days. *edit* attached diagnostics zip, although i believe more focused troubleshooting might be needed? added some ps -elf captures. storage-diagnostics-20150809-1033.zip

just wanted to pop in - previos V5 user. having discovered CA just a few minutes earlier - i am AMAZED at how awesome it makes unraid. actually it is tempting me to go an bug limetech for a secondary usb key just to allow me to play around Keep up the good work Squid and guys!! ps i had a DOH!! moment with my first addon begin kodi-headless. clicking create on chrome gave no result. only when i moved to firefox did i see a small notification saying that more details is needed before i can add. Clicked on advanced settings, and voila - i could add the parameters as requested by the docker. *happy days*

hi guys sorry to hijack - but i am planning to do the same. want to convert a trusty old download pc over to a KVM. in this scenario the windows including download softwares (sickbeard etc etc) s using about 40GB, with the physical drive being 2TB. is it possible to do the raw conversion for the used data only? i don't want to sacrifice a full 2TB of space on my unraid machine, especially since 90% of that wont be used. any ideas? my one thought i had now while typing, is if the raw is not capable of creating a "thin" image, is to use gparted on the original machine to change the partition size - but i would like to avoid meddling with the working hardware.

Hi guys I am having a challenge with s3_sleep (set to shutdown rather than sleep) it seems like all the timers runs out, but then no shutdown happens. Part of the syslog is captured as per below : is there maybe something i am missing? eg do i maybe need to add a powerdown package still? thx Neo_x *Edit* managed to find the answer from earlier in this thread : The above settings act like an unconditional shutdown, in other words your system goes down after 30 minutes - if all works (better define at least one condition). s3_sleep makes use of 'powerdown' to do a proper shutdown of the system, however this is not standard in unRAID and the powerdown plugin needs to be installed. See this topic: http://lime-technology.com/forum/index.php?topic=31735.0 To install powerdown, use the install page under the plugin manager. Copy and paste this URL to install the plugin https://github.com/dlandon/unraid-snap/raw/master/powerdown-x86_64.plg

not sure how to start safe mode? is this possible without having to connect a screen/keyboard? it happened twice again now (eg stock unraid 6.0.1 with only nerdpack added). stayed up for about 12 hours yesterday, as it was streaming movies to the family. this morning it is down again, and even worse - the NFS copy running under MC has crashed/stuck as well thus: GUI down SMB down telnet - Up list user share under telnet - fail ps -ef http://pastebin.com/M2Gvw1wm Top http://pastebin.com/Kd16pcDF lsof -Pni http://pastebin.com/Xq6bfVPf diagnostics capture : http://filebin.ca/27ep8OtR0vNg i managed to perform a diagnost'>http://filebin.ca/27epLpGa0mPY i managed to perform a diagnostics capture, although during capture it reports : didnt have any of this under v5 - and i am running as close to a stock system as possible - no dockers/packages etc ( although nerdpack seems a bit bloated) :'( i will attempt starting safe mode tonight - as connecting screens etc is a bit challenging. *edit - seems possible with editing syslinux * will this syslinux cause safemode? i have updated my go file as per your signature, although there was no additional go file lines. *edit2* ok safemode seems to be going well as per above. screen etc is not working, and only packages installed shows only dynamix will leave this running for a day or two

Hi guys This is quite an odd one, and i hope someone can shed some light. I followed the Limetech upgrade package yesterday as per : https://lime-technology.com/forum/index.php?topic=41061.0 Before starting that, i went into my go file, and disabled all additional packages(mostly just a fan script and unmenu, and rebooted the server. upgrade went without any hitches (and i noticed that the upgrade actually moves all old files into a backup folder, which basically gives me a blank copy to start with.) having prepared a drive before the time, i then started the conversion to xfs (eg formatting a blank drive to XFS, and then copying a full disk to the blank disk - this is/was still going( ETA about 30 hours for 4TB) (i was using screen and mc to accomplish this) my issue is, after about 14 hours, i lost connectivity to the HTTP interface as well as SMB (eg i couldn't start a movie via Kodi) telnet access still worked, after which i tried to reboot the server (sudo reboot) as well as powerdown - both failed - after which i gave it a hard reset (this came up with a parity check which i opted to cancel). I then resumed the XFS migration -all is wel with http access again. this morning, about 8 hours later, the issue repeated - thus i would like to troubleshoot possible causes and get assistance on how to correct this? ( note reverting to V5 is a bit of a mission due to the one XFS drive) ps -a capture http://pastebin.com/4D7Y98ji syslog capture http://pastebin.com/1skRBstL Diagnostics capture http://filebin.ca/27ZlpkrT5H3W i have played around with some of the dynamix plugins, but opted to remove them afterwards to insure some stability during the xfs migration. only package that should be active is nerdpack (to allow me to use screen) any recommendations please? regards Neo_x PS even though the emhttp and SMB issues is active, mc is still copying happily between the [souce] and [dest]/t *edit* added the diagnostics capture. Currently doing a new v6 install from scratch (only restoring disck / network / ident and share config files. will report back if issue occurs again..

Awesomeness!! thank you for your efforts into this - experienced exactly the same. Will implement the two changes , and report back if anything is still going wrong.

Hi guys i had a random case where powerdown caused a parity check afterwards, and i was wondering if there could is someone that could shed some light on the matter? running powerdown v2.06 on a stock unraid 5.0.5 without dynamix. only "addon" is unmenu, which i can probably stop as well if needed. Dont have any APC packages installed. syslog is attached (which was generated by powerdown). Not that the auto_s3_sleep did show all drives was spinned down for 5 minutes before the powerdown was intiated. note since its version 5, no VM's / Xen / Dockers etc is running - just a clean cache drive doing just caching thx Regards Neo_x syslog-20140811-230820.zip

lol same issue was having permissions issues to copy some files via SMB to the array, and thus clicked new permissions, after which it got stuck on the cache drive requiring a hard reset. anyhow long story short. thx for the advice guys - re-adding all the docker images now. just takes forever to download though