Syco54645

No problem. I wish you the best of luck. Before I did the conversion I was super anxious but after I started it really wasn't that bad. Just takes forever.

There was no issue leaving them "unbalanced". I do not have any VM but I would think home assistant will be ok. I am running my Home Assistant in a docker container but want to switch to home assistant os at some point.

If your containers write to the array you'd have to turn those off. Or if you store your docker image on the array. Stuff that only reads and stores data on the cache drive (like Plex) can be left running. You don't need to copy the contents back to the now empty drive. You are missing the formatting of the xfs disk but otherwise your steps seem correct to me.

This I did not know. I have a legacy install that has no value entered, it is just a blue 0 in the box. I will have to go make some changes. Thank you for explaining that.

Thank you. I had somehow missed that. edit: I cannot change this value to 0 for some odd reason. Is there something per device that must be set up as well?

So I am in the midst of creating a new server and I currently have a 1tb disk for the cache pool and a 12tb disk for the array. There is no parity configured at this time. On the Main tab in the UI it shows the disk/array has 1.2tb free yet when I try to move anything else to the array it is reporting no space left on the device. Does anyone have any ideas? indigojr-diagnostics-20240320-2003.zip

I don't really care how it happened just making it known that the setting was apparently enabled automatically on my server. The last time I went into disk settings was to set the default from reiserfs to xfs. I did this when xfs was being pushed over reiserfs. I don't even see how I could have accidentally done this. For me to make a change to the md_write_setting I would have to go into disk settings, change it, and then click save. Regardless. It is solved and that is all that truly matters.

I have not. Speed of writes has never been a concern of mine. Just odd that it was on as I never turned it on, I did not even know this option existed. It has been probably 4 years since I went into a disk settings.

I solved it. Not sure why the md_write_method was set to reconstruct write... I did not do it, I did not know it even existed. I assume an update did and that is slightly concerning.

So I have been having an issue for a while that nearly all of the drives in my array are always spun up. Currently only 1 out of 13 are in standby mode. Yesterday I tried stopping all docker containers and enable them 1 by 1 until it happened. There does not appear to be a correlation there, randomly the drives are just spinning back up. Leaving the containers stopped I saw drives randomly spin up. There was no file access going on during this time. Oddly the read/write counts do not change and the speed is at 0.0B/s for all of the drigves I have tried to use the OpenFiles plugin and the only files being accessed are on disk12 by frigate (issue existed well before frigate was installed), I need to move that off of the array but it is known. Attaching my diagnostics to see if anyone can figure it out. essex-diagnostics-20230411-0823.zip

So I updated the container today and it is failing to boot. Attaching the log below. *** Running /etc/my_init.d/00_regen_ssh_host_keys.sh... *** Running /etc/my_init.d/05_set_the_time.sh... Setting the timezone to : America/New_York *** Running /etc/my_init.d/00_regen_ssh_host_keys.sh... *** Running /etc/my_init.d/05_set_the_time.sh... Setting the timezone to : America/New_York Current default time zone: 'America/New_York' Local time is now: Sun Nov 13 17:59:49 EST 2022. Universal Time is now: Sun Nov 13 22:59:49 UTC 2022. Date: Sun Nov 13 17:59:49 EST 2022 *** Running /etc/my_init.d/10_syslog-ng.init... syslog-ng: Error setting capabilities, capability management disabled; error='Operation not permitted' [2022-11-13T17:59:49.440067] file ../../../glib/gthread-posix.c: line 1339 (g_system_thread_new): error 'Operation not permitted' during 'pthread_create' /etc/my_init.d/10_syslog-ng.init: line 32: 33 Trace/breakpoint trap /usr/sbin/syslog-ng --pidfile "$PIDFILE" -F $SYSLOGNG_OPTS *** /etc/my_init.d/10_syslog-ng.init failed with status 1 *** Killing all processes...

Thanks for confirming that will work. And woops, thanks for pointing that out. That was more of a pseudo command as I am actually doing it per folder on old-disk-12 but this will be useful for someone else if they are trying to do the same things.

Why not just use the old disk and issue the following command. This would avoid having to do a parity resync and fat fingering the new config setup. Not to mention I will maintain parity protection through this process. rsync -rav /mnt/disks/old-disk-12 /mnt/disk12

I did exactly that and had no option to rebuild, it showed only the option to format.

That did not work for me, I had no option to rebuild, only format. The drive is completely empty. It seems the only route forward is copying the data from the old xfs drive over to the array.

I apparently lost data because of this bug. I replaced a 2tb with a 14 tb and now the data is gone. I still have the old 2tb disk and have downgraded to 6.9. What can I do to fix this? Edit: So I thought I would just mount the old drive and write the files back to the new drive... I cannot mount the old drive via usb, I get an invalid super block error. The drive is formatted xfs. If I put the disk back into the system via a sata connection I am able to mount it via the unassigned devices plugin but cant when connected via usb. Any ideas here?

I posted on github after I saw the message to post there instead so will continue there. Posting a reply currently as it looks like all config values are correct. To answer the various failures were caused by not including a callback url. The comment for that line in the json samples is down a line and caused confusion on my end.

Trying to get this setup and working and having issues with the last.fm client. It is telling me "Status: Auth Interaction Required". When I click the link to (re)authenticate or initialize I get various forms of failure. I have created an API key and put the values in to the config.json. I am unsure of how it should function when I get passed to lastfm and then back to the server http://tower.local:9078//lastfm/callback?state=FrankLFM&token=XXXXXXXXXXXX At this point the page says OK and that is it. Heading back to the application it still says that last.fm is not authenticated. Any idea what I am missing? 2021-09-28T09:33:10-04:00 info : [App ] Server started at http://localhost:9078 2021-09-28T09:33:10-04:00 info : [Sources ] (partyPlex) plex source initialized 2021-09-28T09:33:10-04:00 info : [Source - Plex - partyPlex] Initializing with the following filters => Users: N/A | Libraries: party | Servers: N/A 2021-09-28T09:33:10-04:00 info : [Sources ] (FrankPlex) plex source initialized 2021-09-28T09:33:10-04:00 info : [Source - Plex - FrankPlex] Initializing with the following filters => Users: frank | Libraries: N/A | Servers: N/A 2021-09-28T09:33:10-04:00 warn : [Scrobblers ] (FrankLFM) lastfm client auth failed. 2021-09-28T09:33:10-04:00 error : [Client Lastfm - FrankLFM] Error: Invalid session key - Please re-authenticate at CWD/node_modules/lastfm-node-client/lib/ApiRequest.js:136:11 at processTicksAndRejections (internal/process/task_queues.js:95:5) at async LastfmApiClient.callApi (file://CWD/apis/LastfmApiClient.js:84:20) at async LastfmApiClient.testAuth (file://CWD/apis/LastfmApiClient.js:148:30) at async LastfmScrobbler.testAuth (file://CWD/clients/LastfmScrobbler.js:34:27) at async ScrobbleClients.addClient (file://CWD/clients/ScrobbleClients.js:240:27) at async ScrobbleClients.buildClientsFromConfig (file://CWD/clients/ScrobbleClients.js:195:17) at async file://CWD/index.js:128:9 2021-09-28T09:33:10-04:00 error : [Client Lastfm - FrankLFM] Could not successfully communicate with Last.fm API 2021-09-28T09:33:10-04:00 error : [API - Lastfm - FrankLFM] Testing auth failed 2021-09-28T09:33:10-04:00 info : [Scrobblers] (FrankLFM) lastfm client initialized

Thanks! This greatly simplifies things for me. Also thanks trurl for helping as well.

In that case the process seems to be mv /mnt/diskX/* /mnt/diskY/ Once complete make sure all files are moved Stop array Go to Disk Settings and change the File system type to xfs Start array Format the drive that is needed Would that be a good assessment of what I need to do? Is there a better way to move the files than just mv in a screen session?

What method do you recommend for moving files from one drive to another within the array? I thought doing a "mv /mnt/disk1/* /mnt/disk11/" would not update the array. If I am incorrect then I have been given wrong information and this will be quite an easy process.

So how can I empty a drive and not have any new files being added to it at the same time? I guess that is the only stumbling block at this point.

Yes I am in the process of waiting for an 8tb to add to the array for space and in hopes of starting conversions. What would the process look like for this? I would assume I have to stop anything that writes to the array and then use unbalance to move the files and this is where I get fuzzy. When I move the files are they moved as far as the array is concerned or will I have to rebuild parity?

Sure essex-diagnostics-20210919-1729.zip

Wouldn't using new config and placing a drive in the wrong location cause data loss? I do not trust myself to not mess up somewhere along the line with 10 drives to convert. Yes please if you can give advice I would greatly appreciate it (I am on 6.9.0 if it matters). I have posted on the reddit community multiple times and was told that I basically must use unbalance to move all of the files off of the drive then use new config to change the format of the empty drive. This video was given to me as the basic method, just not following the encryption bit.