Crowd Server 3.0.0 displays "web 1.0" GUI, has exceptions in catalina.out in crowd.log directory

[Zlatko Kajan]

Hello,

A few days ago, my Crowd Server stopped displaying the nice GUI (blue bar across the top with Crowd, Applications, Users, Groups, Directories) and started showing what I call a "Web 1.0" GUI instead and lists things as just hyperlinked text in a bullet point list and says "Welcome to the Crowd administration console".

My JIRA/Confluence/Bitbucket/Bamboo still authenticate correctly for all my users, but there is some things this GUI is not letting me do (I.e. password reset).

I don't know if it is related at all, but in crowd.log directory, the catalina.out file is spamming once/second the following Exception:

PluginSchedulerTask-com.atlassian.analytics.client.upload.S3EventUploader:job INFO [com.amazonaws.http.AmazonHttpClient] Unable to execute HTTP request: btf-analytics.s3.amazon.com java.net.UnknownHostException: btf-analytics.s3.amazonaws.com.

at java.net.InetAddress.getAllByName0(InetAdress.java:1280)

at java.net.InetAddress.getAllByName(InetAdress.java:1192)

at java.net.InetAddress.getAllByName(InetAdress.java:1126)

at org.apache.http.impl.conn.SystemDefaultDnsResolver.resolve(SystemDefaultDnsResolver.java:45)

at org.apache.http.impl.conn.DefaultClientConnectionOperator.resolveHostname(DefaultClientConnectionOperator.java:263)

at org.apache.http.impl.conn.DefaultClientConnectionOperator.openConnection(DefaultClientConnectionOperator.java:162)

org.apache.http.impl.conn.ManagedClientConnectionImpl.open(ManagedClientConnectionImpl.java:326)

at org.apache.http.impl.client.DefaultRequestDirector.tryConnect(DefaultRequestDirector.java:610)

at org.apache.http.impl.client.DefaultRequestDirector.execute(DefaultRequestDirector.java:445)

at org.apache.http.impl.client.AbstractHttpClient.doExecute(AbstractHttpClient.java:835)

at org.apache.http.impl.client.CloseableHttpClient.execute(CloseableHttpClient.java:83)

at org.apache.http.impl.client.CloseableHttpClient.execute(CloseableHttpClient.java:56)

at com.amazonaws.http.AmazonHttpClient.executeHelper(AmazonHttpCient.java:464)

at com.amazonaws.http.AmazonHttpClient.execute(AmazonHttpCient.java:273)

at com.amazonaws.services.s3.AmazonS3Client.invoke(AmazonS3Client.java:3660)

at com.amazonaws.services.s3.AmazonS3Client.putObject(AmazonS3Client.java:1432)

at com.atlassian.analytics.client.s3.AnalyticsS3Client.uploadFilesToS3Bitbucket(AnalysticsS3Client.java:88)

at com.atlassian.analytics.client.upload.S3EventUploader.execute(S3EventUploader.java:81)

at com.atlassian.scheduler.compat.local.LocalOnlyPluginJob.execute(LocalOnlyPluginJob.java:30)

at com.atlassian.sal.core.scheduling.TimerPluginScheduler$PluginTimerTask.run(TimerPluginScheduler.java:89)

at java.util.TimerThread.mainLoop(Timer.java:555)

at java.util.TimerThread.run(Timer.java:505)

Over and over and over....

This machine is on a network that has no access to the Internet, never has, and never will.  Why is it trying to connect to Amazon, how do I turn that off, why did it start doing it, and is that the root cause of my "Web 1.0" GUI issue?

[AnnWorley · Accepted solution]

Hi - It looks like you are running into an issue with gzip compression. There is a workaround on this bug report: Tomcat7 produces corrupted gzip response with newest zlib (1.2.11) library

Quoting one of our Crowd developers:

Regarding the connection to analytics.s3.amazonaws.com, this is how Crowd sends some analytics events that helps us to understand how customers use our products. You may learn more about data we collect here https://www.atlassian.com/legal/privacy-policy (this is for all of Atlassian products).

[Zlatko Kajan]

Interesting, we did not update any libraries recently on that machine so I wonder why it just started doing it.  Going to try the workaround.

[Zlatko Kajan]

Workaround worked, thank you!

Also apparently my admin *had* updated zlib actually.

[Steffen Opel _Utoolity_]

Regarding the connection to analytics.s3.amazonaws.com, it is worth noting that I've been  considering Atlassian's analytics approach for BTF systems remarkable transparent and customer friendly in the past. I've always been asked whether to enable/disable it at configuration time, and the analytics configuration page is easy to understand, including instructive examples of which data is collected, and which fields are anonymized.

Atlassian has recently become slightly more pushy though and now requires administrators to 'Acknowledge changes to the privacy policy', without offering to opt-out of analytics collection right then and there. Regardless, the surrounding explanation is still completely transparent and includes links to the example data and analytics configuration page:

• View an <example> of the type of data we collect.
• Once you click to acknowledge the new Privacy Policy, data collection will be enabled unless you <change your settings>, which you may do at any time.

In other words, users now always need to implicitly opt-in to analytics collection first by acknowledging the privacy policy changes, and can then explicitly disable it right afterwards.

Anyway, given your statement that the system "has no access to the Internet, never has, and never will", I suspect that opting out of analytics collection at configuration time slipped through somehow. You can easily disable it as follows:

• Click administration menu 'Analytics'
• Set 'Atlassian Analytics' to 'Disabled'

[Zlatko Kajan]

Yes, this network is not legally allowed to touch the Internet, so....

The "web 1.0" GUI did let me get to 'Analytics' and I selected "Disabled" and clicked save, and that gave me a

HTTP Status 405 - HTTP method Post is not supported by this URL.

type Status report

message HTTP method POST is not supported by this URL

description The specified HTTP method is not allowed for the requested resource

Apache Tomcat/7.0.69

[Zlatko Kajan]

Also, I thought by paying for a Crowd license we get support through the usual way we do for all the other licenses we pay for, I guess the 50 user count license is considered "starter" and we do not?  Is there a 50 user one that is not starter, or do we need a higher user count solely to get the usual support?

[Zlatko Kajan]

Fixing the "web 1.0" issue per workaround above, let me also disable the analytic.

Thank you!

[AnnWorley]

@Zlatko Kajan It is a little misleading that our article about the starter license support offering is entitled We’re updating our support offering for server starter (10-user) licenses Per the table on this page, 50-user Crowd licenses are starter licenses: Starter Licensing and Pricing

It does look like you need to buy the next license tier to qualify for support through our portal, from We’re updating our support offering for server starter (10-user) licenses:

However, If your team requires full-time Atlassian technical support you can upgrade to the next user tier license. See our how-to documentation for instructions.