IC2 errors when trying to retrieve metadata

[Doug Ayers]

This and last week I’ve encountered an issue when retrieving metadata from a sandbox where the “Listing metadata objects” dialog takes ~15 minutes to open.

‌

In the attached log file, I see a SOAP error when trying to retrieve some metadata.

‌

2019-09-18 11:10:03,055 [7755414]   INFO - cloud.client.ForceComApiClient - Logging into Salesforce DX as [e-mail]5. 
2019-09-18 11:10:03,056 [7755415]   INFO - cloud.client.ForceComApiClient - Retrieving org details from sfdx CLI. 
2019-09-18 11:10:03,057 [7755416]   INFO - cloud.client.ForceComApiClient - Retrieving org details from sfdx CLI. 
2019-09-18 11:10:03,057 [7755416]   INFO - cloud.client.ForceComApiClient - Retrieving org details for gus-trailh5 from sfdx CLI. 
2019-09-18 11:10:35,134 [7787493]   INFO - y.ReflectionServiceFactoryBean - Creating Service {urn:partner.soap.sforce.com}PartnerService from class com.sforce.soap.partner.PartnerApi 
2019-09-18 11:10:35,178 [7787537]   INFO - y.ReflectionServiceFactoryBean - Creating Service {http://soap.sforce.com/2006/04/metadata}MetadataService from class com.sforce.soap.metadata.MetadataApi 
2019-09-18 11:10:42,593 [7794952]   WARN - ij.builder.ForceComBuilderUtil - Failed to list metadata as a batch. Will try again one at a time. 
javax.xml.ws.WebServiceException: org.apache.cxf.binding.soap.SoapFault: INVALID_TYPE: sObject type 'sfcertssor__AssetForce_Response_Control__mdt' is not supported.
    at org.apache.cxf.jaxws.JaxWsClientProxy.mapException(JaxWsClientProxy.java:193)

‌

[Scott Wells]

Doug, can you add the following to Help>Debug Log Settings, reproduce this behavior, and provide the resulting log file(s)?

#com.illuminatedcloud.intellij.builder.ForceComBuilderUtil

Also, what type of org is this? And importantly, is it a pre-release org?

[Doug Ayers]

Hi Scott, please see the new attachment with the debug logs per the debug log setting.

This is a pre-release sandbox org.

[Scott Wells]

Thanks, Doug. I'm pretty sure this is a pre-release org issue. There have been quite a few with third-party tooling due to some incompatible SOAP API changes. I'm seeing quite a few errors like this:

2019-09-18 14:18:29,929 [ 675693]  ERROR - ij.builder.ForceComBuilderUtil - Failed to list metadata for ManagedTopic 
javax.xml.ws.WebServiceException: org.apache.cxf.binding.soap.SoapFault: UNKNOWN_EXCEPTION: invalid parameter value
    at org.apache.cxf.jaxws.JaxWsClientProxy.mapException(JaxWsClientProxy.java:193)
    at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
    at com.sun.proxy.$Proxy173.listMetadata(Unknown Source)

that look similar to what I've seen on the discussion group with other tooling partners. These have already been reported to Nate and others, and at least one issue has already been fixed. However, just based on the provided log it looks like there are still some issues using an API v46.0 SOAP client against a pre-release org. I'll need to follow up on that group and see what I can find out about issues like these.

Is the REALLY long OST generation from the other issue also against this org? If so, it may be experiencing some of the same issues.

[Scott Wells]

Yep, at least one of these has already been reported as "V47 bug: CustomMetadata in installed packages causing Metadata API error". That corresponds to all of the failures like the following:

2019-09-18 14:18:25,110 [ 670874]  ERROR - ij.builder.ForceComBuilderUtil - Failed to list metadata for CustomMetadata 
javax.xml.ws.WebServiceException: org.apache.cxf.binding.soap.SoapFault: INVALID_TYPE: sObject type 'sfcertssor__AssetForce_Response_Control__mdt' is not supported.
    at org.apache.cxf.jaxws.JaxWsClientProxy.mapException(JaxWsClientProxy.java:193)
    at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
    at com.sun.proxy.$Proxy173.listMetadata(Unknown Source)

Note that each such failure causes IC to query these again individually vs. in groups to ensure as much metadata coverage as possible as efficiently as possible...but it doesn't anticipate an unexpected error like this which can result in additional API calls which are pretty much guaranteed to fail.

There's another error for an EOF reading the response to the listMetadata SOAP API call:

org.apache.cxf.interceptor.Fault: Unmarshalling Error: Premature EOF 
    at org.apache.cxf.jaxb.JAXBEncoderDecoder.unmarshall(JAXBEncoderDecoder.java:938)
    ...
    at com.sun.proxy.$Proxy173.listMetadata(Unknown Source)

I've just reported that one to the tooling folks as well.

My guess is that until some of these are properly resolved, we won't know for sure whether there's also a performance degradation in the listMetadata API call in Winter '20. Having said that, right now it looks like there may be one.

I'll keep you posted on what else I find, but right now it seems like these are primarily new release teething pains.

[Doug Ayers]

Thanks, Scott, for taking time to dig in and follow up with the right folks!

‌

To your earlier question,

Is the REALLY long OST generation from the other issue also against this org?

Yes, it’s the same org.

[Scott Wells]

Gotcha. I just created a Winter '20 scratch org and went through a full cycle of installing a few managed packages, pushing metadata, and generating the OST. I didn't see any issue along the way, but these managed packages don't contain custom metadata types, nor is the project metadata all that complex. I was hoping I'd be able to reproduce some of this so that I could have it all under my control...at least in terms of being able to provide diagnostics to the folks who REALLY have it under their control! But alas, what I'm currently seeing looks pretty good.

I'll keep you posted on what I hear from the tooling partner folks. Could be that they'll want access to this sandbox org so they can see what's going on with these failing API calls.

[Chris Tufts]

hey Scott,

I have noticed this issue as well… I will see if I can dig up logs if it helps the next time I run into it

[Scott Wells]

Anything new to report here? I haven't heard from anyone on this in a bit.

[Scott Wells]

Resolving this due to lack of activity. If it's still happening, please reopen with new details.