we have been using MLFlow 0.8.2 (with a local file store) for a while, and I was happy to see the release of MLFlow 0.9. After upgrading to the new version, I realized that pointing the MLFLow server to the old file store leads to a non-working web UI (I just see some image of a waterfall).
Is there a recommendation for proper migration of data when upgrading?
Thanks a lot in advance,
Da
Related
Recently been trying to learn Airflow, but a majority of resources online depended on this repo https://github.com/puckel/docker-airflow which unfortunately has been removed.
I am not familiar with docker so I'm just trying to set up locally and play around with Airflow. I'm on a windows setup and have already gotten docker working on my computer. Does Airflow have a quick-set-up file for a docker-compose? Or is there any other resources I can look at? Thanks.
Its a duplicate question.
Use official official docker-compose.yml see here
I recently added a quick start guides to the official Apache Airflow documentation. Unfortunately, this guide has not been released yet. It will be released in Airflow 2.0.1.
For now, you can use the development version, and when a stable version is released it will be very easy for you to migrate. I don't expect any major changes to our docker-compose.yaml files.
http://apache-airflow-docs.s3-website.eu-central-1.amazonaws.com/docs/apache-airflow/latest/start/docker.html
I need to upgrade my solr search from 4.7 version to 5.3.1 .
I am working on a linux platform.
Can you please provide me the steps that i need to follow .
Thank you!
I do not think that there is a definitive step by step guide for the upgrade that you are looking for.
I have a 4.3.x SOLR running in a production environment and I am contemplating the leap to upgrade to 5.x. However its clear that a lot has changed and that my upgrade is not going to be straight forward.
Also other priorities in my project have kept me from doing the upgrade.
So rest of the discussion is more a thought process than actual upgrade experience.
Last I researched I found the below links useful
https://support.lucidworks.com/hc/en-us/articles/203776523-How-to-upgrade-between-major-Solr-Versions
https://cwiki.apache.org/confluence/display/solr/Major+Changes+from+Solr+4+to+Solr+5
From the Major changes link you will notice that a lot has changed ..
Most notably there are changes to the index format, SolrJ removal of deprecated API and that the deployment is now as a standalone server instead of a war file.
So I would suggest that you ask yourself the following questions ...
Is it possible to recreate the index from scratch ? How much time does it take to create your complete index ? If your index can be recreated quickly then , I would suggest that you do that using 5.x engine on a separate machine, while your production environment is served by your existing server. Then plan a complete upgrade from 4.x to 5.x by simply pointing your Production instance to the new SOLR engine. This approach will give you a clean slate to start with and a brand new index (but with existing data).
If you have a very large index (e.g. it takes several days to recreate it from scratch), then you may want to perform an upgrade of the live index. In that case I suggest that you consider the following.
The SOLR upgrade guide mentions 4.10 as a version that is 4.x (so I assume its is easy to upgrade from any 4.x to 4.10) and has some features built in to help with the move to 5.x. So first upgrade to 4.10 ensure that your index continues to work properly. Then use the guides mentioned above to upgrade to 5.x
I've been trying to work through an issue that developed while attempting to upgrade our testing environment from 12.04 to 14.04 ubuntu on aws. Prior to this, the Package repository version of solr was 1.4.1 which matched our 1.4.1 solrj client integrated with our application.
Changing the base AMI to the 14.04 latest and running our default deploy caused solr 3.6.2 server to be installed. It appears it was accepting our configs without issue, however when our client tried to connect we received different errors:
The first was an unknown custom field, which we traced back to our deployment scripts not moving our schema.xml and solrconfig.xml to /etc/solr/conf/ but keeping it in the base directory.
We corrected this issue, and then ran into the following:
'exception: Invalid version or the data in not in 'javabin' format'
This was generated by a wrapper ontop of solrj, but I'll be honest and say I know nothing regarding Solr and that this may be on our end. I've asked our dev team to look at 2 options:
1) enabling: 'server.setParser(new XMLResponseParser());'
Which is the recommendation on the backwards compatibility for an older client.
2) updating our client in the application to 3.6.2
-I know less about the requirements on this.
My fall back is to revert to 1.4.1, but it appears it hasn't been touched since 2011, which makes me hesitant.
Any thoughts / suggestions would be appreciated!
Thanks!
I think the best option is to maintain the same version of Solr and Solrj.
I used for a lot of time Solr 1.4.1 and, while as you said, the most part of it works with newer versions without any problem, actually a lot of things have been changed since 1.4.*
I did your same porting last year, (from 1.4.1 to 3.6.1) and I can confirm you that the 2nd way is the right one: all changes you must do in your client code are just "formal" and very very quick.
Any workaround you could do for being able to communicate with a different version (between Solrj and Solr) is just, as the word says, a "workaround" and it could lead to unexpected (hidden) side-effects later.
I'm using the official nuget package of Windows Azure Storage Client Library to retrieve items of my Azure tables.
Recently I updated the package from version 2.0.2.0 to 2.0.5.0 and my app stopped working because the results returned by my storage query are different with the new version.
I'm looking for the library changelog in order to understand how to fix the issue.
Do you know where can I find it ?
The link provided on the nuget page seems to be outdated (it's a changelog between 1.x and 2.x, not between 2.0.2 and 2.0.5 !). Also, the Windows Azure Storage team's blog is not updated.
Please refer to the changelog.txt that is always updated with the respective source code changes.
The changelog is always up to date at https://github.com/Azure/azure-storage-net/blob/master/changelog.txt
I was recommended by someone in Stack Overflow to use Core Data Editor http://christian-kienle.de/CoreDataEditor/ to manage the sqlite persistent store. However the latest version (3.0) crashes on launch everytime. Older versions load but I see nothing when i point the config to the persistent store and the object model directories. There is no documentation either.
can someone point me to the right place to sort this problem? I am trying to find a more manageable way to coordinate core data development than sqlite consoles.
thank you
You have to set the compiled model in Core Data Editor (ending in .moc, not the .xcdatamodel). You can find it in the resources folder of your app bundle.