upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
Shiori
86 Posts
December 26, 2020, 5:49 amQuote from Shiori on December 26, 2020, 5:49 amupgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
this happened after a full cluster reboot after the upgrade, prior to that everything was fine.
I cant join the cluster as the cluster is up and healthy.
how do I force petasan to use the current config as found in /opt/petasan/ on all mons. the correct settings are there just being ignored and waiting for the node to be commissioned.
I checked with ceph and it says its working and the cluster and pools are still there
upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
this happened after a full cluster reboot after the upgrade, prior to that everything was fine.
I cant join the cluster as the cluster is up and healthy.
how do I force petasan to use the current config as found in /opt/petasan/ on all mons. the correct settings are there just being ignored and waiting for the node to be commissioned.
I checked with ceph and it says its working and the cluster and pools are still there
admin
2,930 Posts
December 26, 2020, 8:10 amQuote from admin on December 26, 2020, 8:10 amupgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
I cant join the cluster as the cluster is up and healthy.
Can you access the Management ui at https port 443 ( new port ) or old port ( 5000 which will redirect to new port ) ? Note 5001 is the Deployment Wizard ui used for new nodes you add and want to join. Is the issue with new nodes you want to install or is with managing existing nodes ?
upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
I cant join the cluster as the cluster is up and healthy.
Can you access the Management ui at https port 443 ( new port ) or old port ( 5000 which will redirect to new port ) ? Note 5001 is the Deployment Wizard ui used for new nodes you add and want to join. Is the issue with new nodes you want to install or is with managing existing nodes ?
Shiori
86 Posts
December 26, 2020, 1:43 pmQuote from Shiori on December 26, 2020, 1:43 pmthis is an upgrade on an existing cluster, if it was a new setup then I would have just installed 2.7.0 directly
the deployment wizard is the only web page being offered, I did a port scan and there is port 80 and 5001. No ports 443 or 5000. This is what I am confused about as it was a simple upgrade so why did the web server state get changed back to not deployed?
this is an upgrade on an existing cluster, if it was a new setup then I would have just installed 2.7.0 directly
the deployment wizard is the only web page being offered, I did a port scan and there is port 80 and 5001. No ports 443 or 5000. This is what I am confused about as it was a simple upgrade so why did the web server state get changed back to not deployed?
Last edited on December 26, 2020, 2:11 pm by Shiori · #3
Shiori
86 Posts
December 26, 2020, 2:51 pmQuote from Shiori on December 26, 2020, 2:51 pmI fixed (sort of) the issue. I reinstalled the Mon nodes one at a time and re-joined the cluster.
I could see that the cluster was up and healthy in ceph and rbd showed the images intact and accessible, so I thought that the only thing missing was something that was either overwritten or wiped out during the upgrade. Not sure why it happened this time as I have two other clusters that were upgraded from 2.3.0 to 2.3.1 to 2.7.0 without issues including my test cluster of odd old parts which dies just by looking at it wrong.
now to drain each storage node and reinstall them as their web pages show not deployed too.
I fixed (sort of) the issue. I reinstalled the Mon nodes one at a time and re-joined the cluster.
I could see that the cluster was up and healthy in ceph and rbd showed the images intact and accessible, so I thought that the only thing missing was something that was either overwritten or wiped out during the upgrade. Not sure why it happened this time as I have two other clusters that were upgraded from 2.3.0 to 2.3.1 to 2.7.0 without issues including my test cluster of odd old parts which dies just by looking at it wrong.
now to drain each storage node and reinstall them as their web pages show not deployed too.
upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
Shiori
86 Posts
Quote from Shiori on December 26, 2020, 5:49 amupgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
this happened after a full cluster reboot after the upgrade, prior to that everything was fine.
I cant join the cluster as the cluster is up and healthy.
how do I force petasan to use the current config as found in /opt/petasan/ on all mons. the correct settings are there just being ignored and waiting for the node to be commissioned.
I checked with ceph and it says its working and the cluster and pools are still there
upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
this happened after a full cluster reboot after the upgrade, prior to that everything was fine.
I cant join the cluster as the cluster is up and healthy.
how do I force petasan to use the current config as found in /opt/petasan/ on all mons. the correct settings are there just being ignored and waiting for the node to be commissioned.
I checked with ceph and it says its working and the cluster and pools are still there
admin
2,930 Posts
Quote from admin on December 26, 2020, 8:10 amupgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
I cant join the cluster as the cluster is up and healthy.Can you access the Management ui at https port 443 ( new port ) or old port ( 5000 which will redirect to new port ) ? Note 5001 is the Deployment Wizard ui used for new nodes you add and want to join. Is the issue with new nodes you want to install or is with managing existing nodes ?
upgraded 2.3.1 to 2.7.0 using the guide and now my mons web page is the install/join webpage on port 5001
I cant join the cluster as the cluster is up and healthy.
Can you access the Management ui at https port 443 ( new port ) or old port ( 5000 which will redirect to new port ) ? Note 5001 is the Deployment Wizard ui used for new nodes you add and want to join. Is the issue with new nodes you want to install or is with managing existing nodes ?
Shiori
86 Posts
Quote from Shiori on December 26, 2020, 1:43 pmthis is an upgrade on an existing cluster, if it was a new setup then I would have just installed 2.7.0 directly
the deployment wizard is the only web page being offered, I did a port scan and there is port 80 and 5001. No ports 443 or 5000. This is what I am confused about as it was a simple upgrade so why did the web server state get changed back to not deployed?
this is an upgrade on an existing cluster, if it was a new setup then I would have just installed 2.7.0 directly
the deployment wizard is the only web page being offered, I did a port scan and there is port 80 and 5001. No ports 443 or 5000. This is what I am confused about as it was a simple upgrade so why did the web server state get changed back to not deployed?
Shiori
86 Posts
Quote from Shiori on December 26, 2020, 2:51 pmI fixed (sort of) the issue. I reinstalled the Mon nodes one at a time and re-joined the cluster.
I could see that the cluster was up and healthy in ceph and rbd showed the images intact and accessible, so I thought that the only thing missing was something that was either overwritten or wiped out during the upgrade. Not sure why it happened this time as I have two other clusters that were upgraded from 2.3.0 to 2.3.1 to 2.7.0 without issues including my test cluster of odd old parts which dies just by looking at it wrong.
now to drain each storage node and reinstall them as their web pages show not deployed too.
I fixed (sort of) the issue. I reinstalled the Mon nodes one at a time and re-joined the cluster.
I could see that the cluster was up and healthy in ceph and rbd showed the images intact and accessible, so I thought that the only thing missing was something that was either overwritten or wiped out during the upgrade. Not sure why it happened this time as I have two other clusters that were upgraded from 2.3.0 to 2.3.1 to 2.7.0 without issues including my test cluster of odd old parts which dies just by looking at it wrong.
now to drain each storage node and reinstall them as their web pages show not deployed too.