May 12, 201511 yr ok so i am currently seeing some "funnies" with dockerhub and thought it might be useful to dev's to have a thread where we can notify each other if there is anything unexpected happening on dockerhub during push/pull of images. my current experience with dockerhub in the last few days has resulted in two issues, which at this time are not resolved:- 1. docker pull down via docker ui in unraid is reporting some seriously whacky download percentages and sizes, this might be dockerhub or client and/or unraid ui related, not sure at this time:- IMAGE ID [cc860c7c90ca]: Pulling metadata. Pulling fs layer. Downloading -3200% of -0MB. Download complete. IMAGE ID [e8425427f88d]: Pulling metadata. Pulling fs layer. Downloading -24800% of -0MB. Download complete. IMAGE ID [ea57f814ed5b]: Pulling metadata. Pulling fs layer. Downloading -72700% of -0MB. Download complete. IMAGE ID [f5dba8b23239]: Pulling metadata. Pulling fs layer. Downloading -4352023900% of -0MB. 2. dockerhub is not updating last build time, so i can see the build time and date stamped correctly for the build but the summary screen is not updating the last build time, this has a nasty side effect that the unraid docker ui is not showing updates available for pulled down images, and there is no way to force an update in the docker ui for unraid.
May 12, 201511 yr I had some problems last night with building containers automatically from github. Deleted my docker repo and tried re-adding them, but it didn't work.
May 12, 201511 yr Author http://lime-technology.com/forum/index.php?topic=39812.0 ahh thanks SB, nice to know somebody else is suffering :-), lets hope it gets sorted soon eh
May 12, 201511 yr http://lime-technology.com/forum/index.php?topic=39812.0 ahh thanks SB, nice to know somebody else is suffering :-), lets hope it gets sorted soon eh absolutely, it's making developing containers hard work.
May 12, 201511 yr Author grrr github is down now, not having a good dev "cloud" day :-) oo its back, just a blip
May 12, 201511 yr grrr github is down now, not having a good dev "cloud" day :-) oo its back, just a blip lol, luckily the container i'm working on today i'm still doing all the local stuff (musicbrainz server) and don't need git yet. might need to pick your brains on something later though, will see how it pans out first.
May 12, 201511 yr Author might need to pick your brains on something later though, will see how it pans out first. no worries, always happy to help out.
May 13, 201511 yr ok so i am currently seeing some "funnies" with dockerhub and thought it might be useful to dev's to have a thread where we can notify each other if there is anything unexpected happening on dockerhub during push/pull of images. my current experience with dockerhub in the last few days has resulted in two issues, which at this time are not resolved:- 1. docker pull down via docker ui in unraid is reporting some seriously whacky download percentages and sizes, this might be dockerhub or client and/or unraid ui related, not sure at this time:- IMAGE ID [cc860c7c90ca]: Pulling metadata. Pulling fs layer. Downloading -3200% of -0MB. Download complete. IMAGE ID [e8425427f88d]: Pulling metadata. Pulling fs layer. Downloading -24800% of -0MB. Download complete. IMAGE ID [ea57f814ed5b]: Pulling metadata. Pulling fs layer. Downloading -72700% of -0MB. Download complete. IMAGE ID [f5dba8b23239]: Pulling metadata. Pulling fs layer. Downloading -4352023900% of -0MB. 2. dockerhub is not updating last build time, so i can see the build time and date stamped correctly for the build but the summary screen is not updating the last build time, this has a nasty side effect that the unraid docker ui is not showing updates available for pulled down images, and there is no way to force an update in the docker ui for unraid. Still getting the weird download today.
May 13, 201511 yr Out of curiosity, is it only errors with dockers from free accounts or are the paid accounts having issues to? I know they have different benefit levels so I wonder if its just the public freebie level that is overloaded or their entire service.
May 14, 201511 yr might need to pick your brains on something later though, will see how it pans out first. no worries, always happy to help out. about that help, lol....... i have the following script and accompanying supervisord.conf file cat <<'EOT' > /etc/my_init.d/01-postgres-setup.sh #!/bin/bash mkdir -p /data/main mkdir -p /var/run/postgresql/9.4-main.pg_stat_tmp chown -R postgres:postgres /var/run/postgresql/9.4-main.pg_stat_tmp if [ -d "/data/main/global" ]; then echo "postgres seems to be configured, will attempt start" chown postgres /data/* chgrp postgres /data/* chmod 700 /data/main exec /usr/bin/supervisord -c /root/supervisord.conf else echo "initialising postgres for external data use" cp /etc/postgresql/9.4/main/postgresql.conf /data/postgresql.conf echo "local all all trust" >> /data/pg_hba.conf sed -i '/^data_directory*/ s|/var/lib/postgresql/9.4/main|/data/main|' /data/postgresql.conf sed -i '/^hba_file*/ s|/etc/postgresql/9.4/main/pg_hba.conf|/data/pg_hba.conf|' /data/postgresql.conf chown postgres /data/* chgrp postgres /data/* chmod 700 /data/main /sbin/setuser postgres /usr/lib/postgresql/9.4/bin/initdb -D /data/main sleep 30s sed -i "/^#listen_addresses/i listen_addresses='*'" /data/postgresql.conf cd /opt/musicbrainz-server /sbin/setuser postgress ./admin/InitDb.pl --createdb --clean sleep 30s exec /usr/bin/supervisord -c /root/supervisord.conf fi EOT cat <<'EOT' > /root/supervisord.conf [supervisord] nodaemon=true [program:postgres] user=postgres command=/usr/lib/postgresql/9.4/bin/postgres -D /data/main -c config_file=/data/postgresql.conf [program:musicbrainz] directory=/opt/music-brainz-server command=plackup -Ilib -r EOT issue i have is , the two commands that have sleep 30s after them , those two processes must both complete before the next operation in the file starts. sleep 30s isn't very elegant and not a guarantee on slower systems that the process has completed, i'm really bad at scripting, any ideas ?
May 15, 201511 yr Author might need to pick your brains on something later though, will see how it pans out first. no worries, always happy to help out. about that help, lol....... i have the following script and accompanying supervisord.conf file cat <<'EOT' > /etc/my_init.d/01-postgres-setup.sh #!/bin/bash mkdir -p /data/main mkdir -p /var/run/postgresql/9.4-main.pg_stat_tmp chown -R postgres:postgres /var/run/postgresql/9.4-main.pg_stat_tmp if [ -d "/data/main/global" ]; then echo "postgres seems to be configured, will attempt start" chown postgres /data/* chgrp postgres /data/* chmod 700 /data/main exec /usr/bin/supervisord -c /root/supervisord.conf else echo "initialising postgres for external data use" cp /etc/postgresql/9.4/main/postgresql.conf /data/postgresql.conf echo "local all all trust" >> /data/pg_hba.conf sed -i '/^data_directory*/ s|/var/lib/postgresql/9.4/main|/data/main|' /data/postgresql.conf sed -i '/^hba_file*/ s|/etc/postgresql/9.4/main/pg_hba.conf|/data/pg_hba.conf|' /data/postgresql.conf chown postgres /data/* chgrp postgres /data/* chmod 700 /data/main /sbin/setuser postgres /usr/lib/postgresql/9.4/bin/initdb -D /data/main sleep 30s sed -i "/^#listen_addresses/i listen_addresses='*'" /data/postgresql.conf cd /opt/musicbrainz-server /sbin/setuser postgress ./admin/InitDb.pl --createdb --clean sleep 30s exec /usr/bin/supervisord -c /root/supervisord.conf fi EOT cat <<'EOT' > /root/supervisord.conf [supervisord] nodaemon=true [program:postgres] user=postgres command=/usr/lib/postgresql/9.4/bin/postgres -D /data/main -c config_file=/data/postgresql.conf [program:musicbrainz] directory=/opt/music-brainz-server command=plackup -Ilib -r EOT issue i have is , the two commands that have sleep 30s after them , those two processes must both complete before the next operation in the file starts. sleep 30s isn't very elegant and not a guarantee on slower systems that the process has completed, i'm really bad at scripting, any ideas ? hi SB, sorry for the delay, work is mental right now, its taking its toll :-). ok so yeah its rather annoying that supervisor doesnt have dependency logic built in, REALLY wish it did, there is a very long thread on google with people +1 this as a new feature, so hopefully it will get added (fingers crossed). in the mean time i am using the following code in a bash script to get around this shortcoming:- # wait for deluge daemon process to start (listen for port) while [[ $(netstat -lnt | awk '$6 == "LISTEN" && $4 ~ ".58846"') == "" ]]; do sleep 0.1 done echo "[info] Starting Deluge webui..." so this is a sample from one of my scripts i use to spin up deluge and openvpn, basically it waits until a specific port is listening, once it is listening it then carries on to start other processes which rely on it. it is a bit rough and ready but actually seems to be fairly robust, maybe you could do something similar?, the alternative to this would be to poll the list of processes running, once its shown then continue, possibly you might need a small delay to allow the process to initialise though, what do you reckon, of use? edit - here is the discussion thread, looks like a startdelay flag did get included into master branch for supervisor, not ideal as it still doesnt monitor that the process is running though :-(, i prefer my hack. https://github.com/Supervisor/supervisor/issues/122
Archived
This topic is now archived and is closed to further replies.