amigenius

Linked dockers like gitlab + postgres + redis cannot be updated / initialized / changed any more if network is set to bridge (older unraid versions with probably older docker versions aren't having this problem). The command /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker create --name='gitlab' --net='bridge'... fails with docker: links are only supported for user-defined networks. This is related to https://github.com/docker/cli/issues/2030 So there are two ways to fix this: 1.) Use this fix https://github.com/pcellix/cli/commit/246a69a04e94332f93a294b073bb233ef9056f6f 2.) Add another option to the Network dropdown 'Default', which omits the --net='bridge' parameter, which is effectively using bridge as network, but doesn't bail out with an error. As long as this is not fixed, the only way to update / initialize / change a docker which is using links is to copy and paste the docker command from the GUI, remove the --net='bridge' parameter and execute it in the shell.

Hi, nope, that's not the reason. /home/git/data is mapped. cya

Hi, these are normal values for Seagates https://sgros.blogspot.de/2013/01/seagate-disk-smart-values.html And i don't think thats the reason. cya

Hi, here are the diagnostics. cya dark-vault-diagnostics-20170509-1234.zip

Hi, I already tried turning off directory cache and shares and dockers. I also installed the file activity plugin, which displayed no activity on the NTSF drives, And just to be sure that i'm not wrong, i reenabled system notifications and guess what happend? poof... all NTFS drives spon up immediatly. I attached the settings which are causing this behaviour. cya

Hi, I think the problem is, that unRaid checks all drives regulary (i think 5 or 10 second interval based on observation of spin ups) when system notifications are on so that unRaid can immediatly send warning notifications via email etc. The problem is, that these checks are somehow incompatible to the UD plugin, causing the drives to spin up. All drives are 4TB Seagate drives mounted as NTFS drives. cya

Hi, they won't spin up when requesting standby status via hdparm. Already tested it. So that's not the reason why they're spinning up when UD and system reports are enabled. And if UD is disabled, they spin down with system reports enabled. So it's cleary a bug in UD. cya

Hi, There is a problem with UD drives and System notifications enabled. If System notifications are enabled, UD drives won't spin down. If i try to spin down an UD drive with hdparm -y /dev/sd* it spins up a few seconds later again. Raid drives do not show this behaviour and will spin down accordingly to the spindown setting. It seems that when System notifications are enabled, there's a regular polling of SMART values like temperature etc. While the RAID drives can handle this accordingly and won't spin up, the UD drives are spinning up--- Please fix it. cya

Hi, I've updated from unRaid 6.1.9 to 6.3.3 and updated also the docker images. Now my ssh keys aren't accepted anymore from the gitlab docker. I've tracked down the issue so far. The problem is the .ssh symbolic link in /home/git. In /var/log/gitlab/supervisor/sshd.log i've found Authentication refused: bad ownership or modes for directory /home/git/data Permissions of the data folder are 777 99:users If i rename the .ssh symlink and create a real .ssh folder and put authorised_keys in it and set all file permissions correctly, ssh auth with public key works again. Problem with this fix is, that when the docker gets updated, i must do this changes again. So how can we fix this better? cya

Hi, many many thanks. added DB_EXTENSION=pg_trgm to the env of the postgresql container and now it works again cya

Hi, my last gitlab update was March 09,2016. cya

Hi, After i've updated redis, postgre and gitlab, gitlab doesn't run anymore. I see in the docker log that it fails at the Migrating database... step The docker just exists I've started the dockers in the correct order redis, postgre and gitlab. please help. cya

Hi, fixed it for the current docker by adding \\\ before the backtick -> \\\` cya

Hi, i found the problem.... My email password contains a backtick ` --link postgresql:postgresql --link redis:redisio ....... --env 'SMTP_PASS=XXXXX`XXX' ...... Thats causing the errors. Somesthing must have changed in one of the last releases, as it worked before this update. cya

Hi, i see this at the end of the update/install cya