By: Vadim Saratovtsev Account Admin 07 Mar 2019 at 10:05 p.m. CST

18 Responses
Vadim Saratovtsev gravatar
With docker edition, to setup cache refresh, we point the oxTrust server by IP, using the Node IP that docker assigns to it. With something like Kubernetes on Docker, if the pod gets destroyed and new one is created, it will have a new IP address. This means that in order for the Cache Refresh to work in the docker/kubernetes environment I would need to update the cache refresh config with a new IP address. The change is not automated. Otherwise cache refresh will stop working. In a regular deployment scenario cluster manager would handle updating that bit when new oxTrust server needs to become a cache refresh master..

By Michael Schwartz staff 07 Mar 2019 at 10:07 p.m. CST

Michael Schwartz gravatar
The point of the IP address in the oxTrust Cache Refresh configuration screen is to make sure that the cache refresh process runs only once. I'm not sure what is the best way to handle this in Docker. Also it's important that the snapshots folder is stored on a persistent volume... so whichever server takes over cache refresh has the latest hash values. Mohammad can look into this for you.

By Mohammad Abudayyeh staff 08 Mar 2019 at 2:08 a.m. CST

Mohammad Abudayyeh gravatar
Hi Vadim, We are looking into it. For the docker / kubernetes enviornments we are figuring out the best way for the cache refresh to be automatically linked when a linkage is broken and recreated. We will get back to you. Thanks,

By Vadim Saratovtsev Account Admin 11 Mar 2019 at 10:14 a.m. CDT

Vadim Saratovtsev gravatar
Thank you Mohammad. Please let me know when you may have time to talk about it more. Michael may have already mentioned it to you.... With best regards, Vadim

By Michael Schwartz staff 11 Mar 2019 at 8:50 p.m. CDT

Michael Schwartz gravatar
I'm closing this issue because there is a similar issue open.

By Vadim Saratovtsev Account Admin 11 Mar 2019 at 10:58 p.m. CDT

Vadim Saratovtsev gravatar
Michael, thats the one opened by you, correct Sir? Trying find a link for it..

By Vadim Saratovtsev Account Admin 12 Mar 2019 at 9:46 a.m. CDT

Vadim Saratovtsev gravatar
We closed the other one, this is the only ticket remaining for Cache refresh high availability.

By Vadim Saratovtsev Account Admin 12 Mar 2019 at 11:04 a.m. CDT

Vadim Saratovtsev gravatar
Mohammad, how are we looking Sir?

By Aliaksandr Samuseu staff 12 Mar 2019 at 12:57 p.m. CDT

Aliaksandr Samuseu gravatar
Hi, Vadim. I think Michael could mean [this ticket](https://support.gluu.org/authentication/6754/cache-refresh-will-not-initialize/#at45988). Here re-opened it, and reassigned to another developer.

By Mohammad Abudayyeh staff 14 Mar 2019 at 10:19 a.m. CDT

Mohammad Abudayyeh gravatar
Hello Vadim, Sorry for this late response we were running tests. For now, you will have to be manually updating Cache Refresh settings. However, we are releasing this feature hopefully in 4.0 . We have opened an issue for it and set a plan on how to implement this feature. If you have no further questions we can close this ticket. Thanks,

By Vadim Saratovtsev Account Admin 14 Mar 2019 at 10:29 a.m. CDT

Vadim Saratovtsev gravatar
Mohammad, Thank you Sir. Not sure if it is a question to Michael or you. Do you have a projected release date estimate for 4.0 I can share with the client? Also, any documentation on the projected functionality enhancements ? Many thanks again. Vadim

By Mohammad Abudayyeh staff 14 Mar 2019 at 12:15 p.m. CDT

Mohammad Abudayyeh gravatar
No problem Vadim. Here is our [road map](https://www.gluu.org/roadmap/) you can view estimated time delivery for our releases . The opened issue related to this ticket is [here](https://github.com/GluuFederation/oxTrust/issues/1566) Thank you . Please inform us if you have any more questions

By Vadim Saratovtsev Account Admin 14 Mar 2019 at 1:47 p.m. CDT

Vadim Saratovtsev gravatar
Mohammad, just so I understand, even with Cluster Manager, we will have to update IP address of the master cache refresh server manually every time there is a failover scenario?

By Mohammad Abudayyeh staff 14 Mar 2019 at 4:11 p.m. CDT

Mohammad Abudayyeh gravatar
Hopefully there will not be any failovers but the current CR setup requires manual input of the IP address and has no way to figure out how to automatically update itself. We will try our best to expedite this feature. Please feel free to check back with us, or directly check the status of the [open issue](https://github.com/GluuFederation/oxTrust/issues/1566).

By Michael Schwartz staff 14 Mar 2019 at 6:24 p.m. CDT

Michael Schwartz gravatar
I will expedite this fix. I don't see it as a feature enhancement.

By Vadim Saratovtsev Account Admin 14 Mar 2019 at 7:38 p.m. CDT

Vadim Saratovtsev gravatar
Thank you !!!

By Mohammad Abudayyeh staff 19 Mar 2019 at 1 p.m. CDT

Mohammad Abudayyeh gravatar
Hi Vadim, We are currently working on a solution. Thanks for being patient. We will update you soon. Thank you

By Vadim Saratovtsev Account Admin 28 Mar 2019 at 1:56 p.m. CDT

Vadim Saratovtsev gravatar
Please close the ticket. Thank you!!

By Mohammad Abudayyeh staff 28 Mar 2019 at 3:48 p.m. CDT

Mohammad Abudayyeh gravatar
Thank you Vadim !