Stream: dev

Topic: Deploying 6.1


view this post on Zulip Don Richards (Sep 13 2024 at 19:22):

I just finished upgrading to 6.0 without any significant issues. I'm now attempting to update to 6.1, but I'm encountering a problem right from the start. Here are the exact steps I'm following:

# Step 0:
export PAYARA=/usr/local/payara6
sudo -u dataverse bash -c "echo 'export PAYARA=/usr/local/payara6' >> ~/.bashrc"
sudo -u dataverse bash -c "echo $PAYARA"
# Step 1:
sudo -u dataverse bash -c "$PAYARA/bin/asadmin undeploy dataverse-6.0"
# Step 2:
sudo systemctl stop payara
sudo rm -rf $PAYARA/glassfish/domains/domain1/generated
# Step 3:
sudo systemctl start payara

Issue: After deleting the generated directory I can't restart payara. Step #3 fails at this point. However, if I redeploy 6.0, it starts without any problems. I haven't made any changes to the codebase (e.g., no git pull or similar operations).

Has anyone encountered this issue or have any suggestions on how to resolve it? Should I skip starting payara?

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 16 2024 at 13:05):

Hmm. My first thought is about systemctl stop working properly. Maybe there's still a Paraya process that needs to be killed?

view this post on Zulip Don Richards (Sep 17 2024 at 12:58):

Our service runs the stop-domain command.

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 17 2024 at 13:17):

I'm not sure what to tell you. I used to blow away the "generated" directory all the time in development and I don't remember Payara not starting afterward. :thinking:

view this post on Zulip Don Richards (Sep 17 2024 at 15:01):

This might be slightly unrelated, but I'm upgrading our cloned Dataverse site from 5.14 to 6.0. We have a full copy of the database for testing, but to save costs, we didn't copy over the data mount (i.e., /usr/local/dvn/data). Will the lack of mounted content cause any issues post-upgrade?

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 17 2024 at 15:18):

Hmm, I wouldn't think so.

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 17 2024 at 15:19):

@Don Sizemore do you have any thoughts on this?

view this post on Zulip Oliver Bertuch (Sep 17 2024 at 15:36):

Is there anything relevant in the server.log that might us here? (It sounds like you are starting Payara again in step 3, so there should be log output why things fail in there)

view this post on Zulip Don Sizemore (Sep 17 2024 at 16:13):

@Don Richards did you get the osgi-cache subdirectory as well? old modules can sometimes cause problems. As Oliver says the answer should be in Payara's server.log.

view this post on Zulip Don Richards (Sep 17 2024 at 17:33):

I'm seeing this a lot

[2024-09-17T10:45:02.382-0400] [Payara 6.2023.8] [WARNING] [] [jakarta.enterprise.web] [tid: _ThreadID=98 _ThreadName=http-thread-pool::jk-connector(4)] [timeMillis: 1726584302382] [levelValue: 900] [[
  StandardWrapperValve[Faces Servlet]: Servlet.service() for servlet Faces Servlet threw exception
java.io.IOException: getDataAccessObject: Unsupported storage method.
...

view this post on Zulip Don Richards (Sep 17 2024 at 17:49):

I noticed it more than one thing:

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 17 2024 at 18:17):

Hmm, do you have a custom metadata block called "software"?

view this post on Zulip Don Sizemore (Sep 17 2024 at 18:23):

@Don Richards on not copying over the data, that's exactly how I test with our staging server. just results in 404s for existing datafiles. on unspported storage method: have you reviewed your dataverse.files.* jvm-options?

view this post on Zulip Don Richards (Sep 17 2024 at 21:28):

I'm not sure I follow.


Last updated: Nov 01 2025 at 14:11 UTC