lankanmon

Okay, I will see if I can update. I also managed to get the docker image to run. I do see a lot of BTRFS related errors created by docker. I have attached the most recent diagnostics, can you please see if these errors are of concern or if there is anything I should do to reduce them? I am worried that there are lasting effects from the cache issues. Thanks! lknserver-diagnostics-20190127-2017.zip

Okay so I went back and looked at the other thread and saw that, if it failed, I had to run the command with a smaller param. So I went back and ran it again (each time) with -dusage=1 -dusage=10 -dusage=25 -dusage=50 -dusage=60 -dusage=70 -dusage=75 Each time, it did succeed. Then after that, I restarted the server and it seems to work now. I would still like to know if I need to use a larger number (greater than 75) and what the limit is. Also how often I must run this command and if there is a way to automate this process. BONUS: What is btrfs scrub status? and do I need to run that as well? I want to thank all of you for your help. I really appreciate it!

I have the command running right now. It appears to be working and am at 50% btrfs balance start -dusage=75 /mnt/cache I just wanted to know what I need to know after this is done... will the drive automatically no longer be read-only? Do I need to restart? Also, I ran it with -dusage=75 Do I need to run it again with a bigger number and is the max 100? I just want to know what I can do to prevent this from happening again. UPDATE: It just finished and I got this error: Jan 27 14:35:27 LKNServer kernel: BTRFS info (device sdf1): relocating block group 370487918592 flags data Jan 27 14:35:33 LKNServer kernel: BTRFS info (device sdf1): found 7445 extents Jan 27 14:36:58 LKNServer kernel: BTRFS info (device sdf1): found 7445 extents Jan 27 14:36:59 LKNServer kernel: BTRFS info (device sdf1): relocating block group 369414176768 flags data Jan 27 14:37:06 LKNServer kernel: BTRFS info (device sdf1): found 6155 extents Jan 27 14:38:54 LKNServer kernel: BTRFS info (device sdf1): found 6155 extents So, what should I do now?

lknserver-diagnostics-20190127-1224.zip

Also, when I start the docker service, I get: Jan 26 19:31:40 LKNServer ool www[3156]: /usr/local/emhttp/plugins/dynamix/scripts/emhttpd_update Jan 26 19:31:40 LKNServer emhttpd: req (15): cmdStatus=apply&csrf_token=**************** Jan 26 19:31:40 LKNServer emhttpd: Starting services... Jan 26 19:31:40 LKNServer emhttpd: shcmd (5104): /usr/local/sbin/mount_image '/mnt/user/system/docker/docker.img' /var/lib/docker 30 Jan 26 19:31:40 LKNServer root: truncate: cannot open '/mnt/cache/system/docker/docker.img' for writing: Read-only file system Jan 26 19:31:40 LKNServer kernel: BTRFS info (device loop3): disk space caching is enabled Jan 26 19:31:40 LKNServer kernel: BTRFS info (device loop3): has skinny extents Jan 26 19:31:40 LKNServer kernel: BTRFS info (device loop3): bdev /dev/loop3 errs: wr 68, rd 0, flush 0, corrupt 0, gen 0 Jan 26 19:31:40 LKNServer kernel: BTRFS warning (device loop3): log replay required on RO media Jan 26 19:31:40 LKNServer root: mount: /var/lib/docker: can't read superblock on /dev/loop3. Jan 26 19:31:40 LKNServer kernel: BTRFS error (device loop3): open_ctree failed Jan 26 19:31:40 LKNServer root: mount error Jan 26 19:31:40 LKNServer emhttpd: shcmd (5104): exit status: 1 What can I do to solve the BTRFS errors? The cache drive is not full and I reduced some files too, but still getting the read-only error.

Okay, so I disabled the docker (the vms were already disabled) and started the server. It finally mounted again. I tried to runt he balancing command, but did did not work... root@LKNServer:~# btrfs balance start -dusage=75 /mnt/cache ERROR: cannot access '/mnt/cache': No such file or directory and the web interface shows: and the Balance button does not seem to work either... I am also going to do what @limetech said lower ( ) by moving everything out of the drive, formatting and putting things back on. I have started the mover to get things off the drive naturally to see if that can do all the work for me. I have temporarily marked the usually cached shares as Yes instead of Prefer so I can move them out of the cache.

Disk 2 is full and I have been looking for ways to distribute the content to other disks, but have not found a way besides doing it manually... I what does it mean that the drive if fully allocated and what can I do to fix this prior to starting array? I am unable to start the array as it stalls. Thanks for the feedback!

Thank you! I have attached the diagnostics file. I must clarify that the server does turn on and I am able to access the web GUI. It's just that when I click Start array, it just shows Array Starting - Starting services... (but does not actually start). Please take a look at the diagnostics file and see if it can tell you something. Thanks! lknserver-diagnostics-20190125-2045.zip

Hi all, I have recently had a power outage due to winter weather and my server shutdown after running on my UPS for a while. I think it got the shutdown command and shut itself down -- the UPS was not depleted and continued to power wifi long after server shutdown. Ne next day, when the power was back and stabilized, I turned it back on and it seemed to work fine, but I noticed that my docker images were not behaving normally. Plex videos would stop working after playing for a while. I checked to see if a parity check was running, but it wasn't. Then I decided to restart the server properly. I turned off my docker service. Took the array offline and then restarted the server with the web button. After restarting this time the web GUI was non-responsive and would not even start docker. It said the docker service has failed. I then tried to restart docker and this time the web GUI became non-responsive. I waited for a few hours and then force restarted the server - the unclean way :( After starting again, it detected the unclean shutdown and started a parity check after starting the array. I then let it run for 24hrs (usually it is much less) and when it checked, the web GUI was unresponsive again and the parity check was at 1% (45GB). I restarted again because, at this time, I need to get it working and was really tired of waiting 3 days. Now on boot... it is stuck on Array Starting - Starting services... I have now been 45min... I don't know what to do and feel helpless... Please advice on what I can do to recover the server. I am running on a headless server, so I can't use the non-web interface. Thank you!

Okay, thanks! I will try that out,

I cleaned my server after about 9 months of continuous operation, during which time, two of my drives became disabled (probably due to wire disconnect). After rebooting, the contents are emulated even though both drives are now detected by the server (the serials appear in the dropdown) The drives messages are: I am positive that both drives became unavailable after this cleaning and booting with the drives not connected (and then rebooting made them appear in the dropdown). When I select the correct drives and start the array, it mounts, but the drives show the above errors. How can I tell the server that it is safe to mount them? The content is currently emulated. Thanks!

+1 Would be very useful for sure. I am currently using Krusader, but it is not the most efficient way to do it.

Interesting... I am sure the Devs will be able to find the best way to handle this. It is a bit beyond my knowledge of how unraid tracks file changes. I just noticed that it is able to tell you that that the parity is invalid when the contents are updated outside of parity. For me, this does not work. Even opening up a new tab still waits for the operation to complete before that page loads (waiting for a connection). I will check that out... Thanks!

I made a post requesting help a few weeks ago about this issue that had no responses. However, I think this feature is beneficial. Something that has been bugging me for a while is the fact that hitting the Compute (or Compute All) button on the shares page will require the server to calculate the storage usage for that share. From my experience, this locks up the web UI while the size of the share on each disk is calculated. This tends to take a long time depending on how big the share is and will make the WebUI unresponsive. This is a massive productivity hit because I am usually only doing a compute of space if I am working on the settings of my server. This forces me to take a break and wait for the server to become responsive again. Request: If it is possible, please change the way that shares are computed, such that the calculation is done in the background (Asynchronously) and the results visible whenever they are done (maybe with a notification to let you know when it is ready). If Implemented: I do realize that calculations being made asynchronously have the potential for inaccuracy as more data may be written to disk in the meantime. Maybe have an Icon to show if the size computed is no longer valid (per drive in the results) for information. I think it would use a similar implementation that is currently used to show the orange triangle for unprotected drives. That way, we can hit computer or computer all without having to worry about the delay that it will cause my workflow. Thank you!

Hi all, I accidentally hit the compute all button in the user shares section while trying to add a share (was distracted ). It locked up my unraid webgui and it is not responding (I am assuming it is trying to index all of the 15 or so TB of data I have on the disks). Is there any way to kill the process via terminal? Thanks!

Hi all, So I have recently been having an issue that others may also be having and that is getting notifications that are may or may not be welcome. First, my specific use case: I have posted about this issue here before, but essentially, ever since I purchased a UPS for my server, I have been randomly getting error messages like this: Time Event Subject Description Importance 22-10-2017 10:48 PM unRAID Server Alert UPS Alert Communications restored with UPS {ServerNameHere} alert This is not a dangerous issue, especially since the message is saying "restored" (ie, positive). But the problem is, it happens often at times and sometimes not at all for two or three days. However, this is not something that is really high risk (at least to me). Now, being the paranoid person that I am, I have setup my server so I will email me whenever I get any error, just so I don't miss something important. This leads to my inbox filling up with server alerts each time this is triggered. Not to mention the minor heart attacks I get every time I go to the server GUI and see the red alert bubble in the top right corner. Request (TL,DR.): All I am asking is for a way for one to mute a specific notification from showing up (or triggering emails). Even if this does it by immediately archiving the alert (so it remains as a record). In addition, a way to reorganize alerts into different priorities or Importance. (<- Wishlist) I have ideas for implementation of this if you are interested, but I would definitely like to know if this issue is something that only I am having or if others would also benefit from muting certain notifications. Thanks for your consideration!

I have not done a safe mode. Mainly because this behavior is erratic. Last week it occurred 3 days in a row, and then stopped and the occurred again last night. Since I do use the server daily, I did not have time to leave it in safe mode, but I will do that if I can. In regards to the logs, from the appearance, it only has logged since reboot (which I will post below). Is there any way to view logs from before? Thanks! Log (syslog.txt): https://pastebin.com/qNzNEQS4 (not sure if I should post 2000+ lines here, so I used paste bin, please let me know if I should post it here)

Hi All, So recently, I have been having some issues with my Unraid server where (usually overnight), it is becoming unresponsive and inaccessible over the network. Usually, when this happens, I am unable to access the WebGUI or any of the Dockers running on it as well as all of the VMs. I am able to connect to it via SSH, however, I am unable to shut it down with the poweroff command. I usually have to manually force a shutdown via my motherboards IPMI and then reboot. Obviously, this is an unsafe shutdown method and require a parity rebuild (with 0 errors) afterward. I have noticed that Unraid is operational but in-accessible (the fact that I can connect via SSH and the that the Desktop Login screen is visible on the Thumbnail in my IPMI web gui). I was hoping that someone knows why this may be. This has been happening more and more recently and I have no way to explain what that cause is. I am not too familiar with reading logs, so if someone can walk me through, that would be much appreciated. Thank you!

So, I have been away for close to a month and returned to my server having some SMART errors. I was just wondering how risky these errors are and whether I need to replace the drive... Any help is much appreciated. Thank you!

I have purchased a UPS last week, specifically: https://www.cyberpowersystems.com/product/ups/cp1500avrlcd/ I have the server plugged in directly to a battery+surge protected outlet on the back of the UPS and have the USB B to USB A cable connected directly to a USB 2.0 header on my SuperMicro Motherboard. It appears to be working fine, but starting about 3 days ago, I have been getting the following alert daily. Time Event Subject Description Importance 22-10-2017 10:48 PM unRAID Server Alert UPS Alert Communications restored with UPS {ServerNameHere} alert Today, I saw that it had occurred 3 times. I just wanted to know why this is occurring, if its a problem, and if not, is there a way to specifically mute this alert? GUI to see a Red Notification only to learn it is this message. Please let me know, Thanks!

Do you use ssh? I have had issues where I used screen through terminal and screen keeps accessing the array keeping it from being unmounted. Look for things that may be prevent unmounting of the array. Also, disable docker in settings and then try. Just turning them off individually may not be enough and maybe try unmounting array on its own before hitting shutdown.

So in general, you can do One drive as Parity and Two or More drives for the Storage (the only limitation is your license or max SATA ports on you motherboard). You do not need more than one parity drive as Unraid is designed specifically to allow you to achieve this. One important thing to note is that the parity drive needs to be EQUAL or GREATER in size than ANY of the storage drives used to allow you to parity. So if you have three 4TB drives, then you can have one 4tb parity and two 4tb storage to give you a total of 8tb of parity-protected storage. In the same sense, if you have two 6tb drives and one 4tb drive, then you must use one 6tb drive as parity and the 4tb and 6tb drives as storage to give you a total of 10tb of parity-protected storage. You only need two parity drives for additional redundancy, in case your parity drive fails. Even in that case, as long as your storage array is intact, all you need to do is replace the drive and rebuild the party. I, myself have 5 disks for storage and 1 parity. In terms of naming and identifying the drives, I made myself an excel spreadsheet with all of my drives with their identification (model and serial numbers) as well as a physical label on my rack. I connected all of the drives at once and booted unraid and selected the respective drives that I wanted. Ideally, you want your newest one to be the parity drive (just because of how often it is used). If you can, let us know your drive config if you have any other questions.

Okay, so I see that there is no possible way to remotely control the desktop UI from another device. What I wanted to do was simply connect a monitor to the servers VGA port and have it open to the browser url for netdata (localhost:19999) or have htop running or to have a terminal open with a screen session. Just whatever server thing I wanted to be displayed on that monitor. But doing more research, it does not seem possible since you cant really install anything to the unRAID os. So what I think I need to do is have a separate VM with GUI and have the display of that output over my VGA and have that display the content. It won't be as efficient but seems like the simplest solution under the circumstances. When I opened the thread, I was hoping there was a way to connect to the Desktop GUI via VNC or to install something like synergy, but that does not seem to be possible (for good reason). Sorry for the confusion folks. Thanks!

Sorry, I don't think I am explaining my question. The GUI that I want to access is the Desktop GUI because I want to leave it displayed on a monitor (so I can see things at a quick glance). I am aware of the Web GUI and use it often for actual unRAID maintenance. I need to access the Desktop GUI only because I want to show server stats on a dedicated monitor.

Just FYI, the bell icon will only show you notifications from topics that you follow by default. At the top of each post, there is a Follow button. You need to click it to be notified of any activity within that thread. There is also little switch at the bottom left of reply boxes that let you be notified of replies to you as well. Hope that helps.