Daily Log for #alfresco IRC Channel

Alfresco discussion and collaboration. Stick around a few hours after asking a question.

Official support for Enterprise subscribers: support.alfresco.com.

Joining the Channel:

Join in the conversation by getting an IRC client and connecting to #alfresco at Freenode. Our you can use the IRC web chat.

More information about the channel is in the wiki.

Getting Help

More help is available in this list of resources.

Daily Log for #alfresco

2019-06-20 11:16:01 GMT <alfresco-discord> <MorganP> When using an Alfresco 6.1 cluster, if I want to have 1 ActiveMQ on each Alfresco Nodes, can they work in standalone? Or do they need to be in an ActiveMQ Cluster as well? (or Master/Slave)

2019-06-20 11:35:43 GMT <AFaust> I assume the ActiveMQ would need to be clustered properly. Otherwise your view on events and what not in the event queue will be incomplete / completely inconsistent.

2019-06-20 11:51:30 GMT <alfresco-discord> <MorganP> The doc isn't really clear about that, it just redirect to the Apache ActiveMQ website...

2019-06-20 11:51:45 GMT <alfresco-discord> <MorganP> I guess I will go with a Master/Slave activeMQ

2019-06-20 13:34:26 GMT *** mmccarthy1 is now known as mmccarthy

2019-06-20 14:38:40 GMT <alfresco-discord> <jpotts> I have an Enterprise client who installed the S3 connector and used S3 storage for everything for quite a while. Now they want to pull everything off of S3...

2019-06-20 14:39:06 GMT <alfresco-discord> <jpotts> Our original plan was to install the content store selector and add the file content store as an option, then write scripts to set the cm:storeName to the file content store on all objects.

2019-06-20 14:39:35 GMT <alfresco-discord> <jpotts> That actually worked fairly well for active objects that are not checked out and are not declared as records.

2019-06-20 14:39:48 GMT <alfresco-discord> <jpotts> Versions and archived objects are a problem.

2019-06-20 14:40:29 GMT <alfresco-discord> <jpotts> For archived objects, I should be able to use the node service to set the content store. I'm hoping Alfresco will see that and migrate the object.

2019-06-20 14:40:47 GMT <alfresco-discord> <jpotts> For versioned objects, I don't think I can add aspects or set properties to objects in the version store.

2019-06-20 14:41:47 GMT <alfresco-discord> <jpotts> I think our options are to either leave versioned objects in S3 or manually migrate the objects and update the DB directly, which I really don't want to do. Can anyone think of any other ideas?

2019-06-20 14:43:20 GMT <angelborroy> Patching Alfresco / S3 connector is not an option, right?

2019-06-20 14:43:57 GMT <alfresco-discord> <jpotts> In what way would the connector need to be patched?

2019-06-20 14:44:07 GMT <angelborroy> I don’t know that connector

2019-06-20 14:44:20 GMT <angelborroy> You said: “Versions are a problem”

2019-06-20 14:44:21 GMT <angelborroy> Why

2019-06-20 14:44:22 GMT <angelborroy> ?

2019-06-20 14:44:26 GMT <angelborroy> They are not migrated?

2019-06-20 14:44:56 GMT <alfresco-discord> <Thijs> @jpotts I can tell you we did some content store migrations on the db directly

2019-06-20 14:45:11 GMT <alfresco-discord> <jpotts> Versioned objects are sitting in the version store with pointers to S3. We want those objects to be in the file content store, not S3. So we'd like to just update the versioned object with the cm:storeName so that Alfresco would handle it

2019-06-20 14:45:12 GMT <alfresco-discord> <Thijs> and yes, you have to be careful with the crc

2019-06-20 14:45:36 GMT <angelborroy> I was thinking on extending Alfresco to provoke that migration on cm:storeName changing

2019-06-20 14:45:44 GMT <angelborroy> But I don’t know if I’m saying something weird

2019-06-20 14:47:04 GMT <alfresco-discord> <jpotts> @angel.borroy If I try to use the NodeService to change a versioned object using its versionStore://version2Store store reference the exception is that the operation is not allowed in the versionStore. The current version of the object can be updated and that gets migrated just fine.

2019-06-20 14:47:19 GMT <alfresco-discord> <jpotts> It is the past versions that are the problem.

2019-06-20 14:47:52 GMT <alfresco-discord> <jpotts> @Thijs Doing it by hitting the DB directly just seems very risky and time-consuming.

2019-06-20 14:48:14 GMT <alfresco-discord> <Thijs> @jpotts Yes, we hit a problem with that, but it was not time-consuming

2019-06-20 14:48:34 GMT <angelborroy> Overwriting NodeService for the migration to update (also) versionStore?

2019-06-20 14:49:24 GMT <alfresco-discord> <Thijs> @jpotts We moved from filecontentstore to something S3 like

2019-06-20 14:50:09 GMT <alfresco-discord> <jpotts> I guess it may come down to how bad the client wants to be completely off S3

2019-06-20 14:51:43 GMT <alfresco-discord> <jpotts> In the future I would advise people to not go 100% S3. Instead, use the content store selector and have certain objects go to S3. This wouldn't help me with the version history problem but it would keep internal things like surf config and tag scopes local.

2019-06-20 14:52:11 GMT <alfresco-discord> <Thijs> I don't say you should, but if you manipulate, make sure you have the CRC right, or you will loose data

2019-06-20 14:52:13 GMT <alfresco-discord> <Thijs> CrcHelper.getStringCrcPair(contentUrl, 12, false, true);

2019-06-20 14:52:39 GMT <alfresco-discord> <jpotts> Thanks, @Thijs, that is a good tip!

2019-06-20 14:52:54 GMT <AFaust> jpotts: Dumb question - why is it not possible to simply move the content off S3 into a simple file based structure (with a tool / client outside of Alfresco), using the content URL as a template for the folder structure, and than have a slightly customised file store in Alfresco resolve S3 content URLs against that file structure. Combine that with an aggregating content store, using this S3 file structure as a fallback, and

2019-06-20 14:52:54 GMT <AFaust> you should be golden, no?

2019-06-20 14:55:23 GMT <alfresco-discord> <jpotts> @AFaust: Not a dumb question at all. This client is very paranoid about customizations. They agonize over even the smallest Share tweaks.

2019-06-20 14:56:05 GMT <alfresco-discord> <jpotts> I suspect they'd pick the "direct DB" option over the "customized file store" option, even though the former is probably riskier.

2019-06-20 14:58:28 GMT <AFaust> But if they pick the "direct DB" option, then they hopefully opt for the "simply change the content URL in alf_content_url table" option instead of putting in extra cm:storeName property for all version objects...

2019-06-20 14:58:48 GMT <alfresco-discord> <jpotts> AFaust: Oh yes, definitely

2019-06-20 15:01:33 GMT <alfresco-discord> <Thijs> @jpotts Is there any encryption in play?

2019-06-20 15:01:47 GMT <alfresco-discord> <jpotts> @Thijs: No, thank goodness

2019-06-20 15:02:45 GMT <alfresco-discord> <Thijs> I don't think it should matter, but I would definitely be less at ease

2019-06-20 15:02:57 GMT <alfresco-discord> <jpotts> One less thing to worry about

2019-06-20 15:03:20 GMT <alfresco-discord> <Thijs> postgres?

2019-06-20 15:03:24 GMT <alfresco-discord> <jpotts> Oracle

2019-06-20 15:25:45 GMT <alfresco-discord> <dgradecak> once Centera stopped supporting XAM, I thought would be a nighmare, but somehow we found a way to recalculate the object id from XAM into the centara object id and did that change directly on the DB and sure we had to implement store for with centera sdk supported

2019-06-20 15:27:06 GMT <alfresco-discord> <dgradecak> I am not sure what kind of storage S3 has but if all of that could be exported like AFaust says in a normal file structure, than that is not a big issue I guess

2019-06-20 15:27:24 GMT <alfresco-discord> <dgradecak> but you are a celever guy 😉

2019-06-20 15:28:13 GMT <alfresco-discord> <jpotts> @dgradecak Getting the files out of S3 is no problem. There are S3 APIs in lots of languages. And good ole REST if needed.

2019-06-20 15:29:07 GMT <alfresco-discord> <dgradecak> so that is the easiest way, now just be persuadive 😄

2019-06-20 15:30:10 GMT <alfresco-discord> <jpotts> I kind of wish we had gone the direct DB route initially. Using the content store selector and running the scripts has taken a lot of time. It would have been much easier just to move the entire S3 bucket to disk, then update the DB.

2019-06-20 15:31:14 GMT <alfresco-discord> <dgradecak> yes, if you can have the relation between the bucket and the exported structure on disk, than it is ok to update with that "algorithm" directly in the db

2019-06-20 15:31:31 GMT <alfresco-discord> <dgradecak> that is hwat I did for centera6-7 years ago

2019-06-20 16:31:25 GMT *** jelly-home is now known as jelly

End of Daily Log

The other logs are at http://esplins.org/hash_alfresco