Gateway timeout in SAP Data Hub 1.0 SPS 3 trail version

[haley_zhang]

Hello experts,

Could you please help with the following issue about Data Hub?

I am facing the error of gateway timeout when opening Data Hub Cockpit with tab Landscape and Discovery. I am using the trail version for 1.0 SPS 3.

This issue has lasted for a long time(more than one week) and I've tried to suspend and reactivate the instance several times but it didn't help at all.

Also, following the note 2602716 - Release Restrictions for SAP Data Hub 1.0 SPS 03,

"Non reachable HANA systems cause Cockpit Overview and Landscape to timeout

When a HANA connection was created, and the HANA instance might not responding (due to error), the Data Hub Cockpit might receive a gateway timeout when calling Data Hub Objectservice by /v1/systems. This behavior does not occur on first creation of HANA connection, but might happen, when HANA DB is not reachable after creation."

I can connect to HANA DB with SYSTEM user on HANA studio during the whole time.

As for the "flow agent" mentioned in other place, it also looks good from my point of view and was not deleted.

I don't want to delete the instance and work with a new one as I've already built some models and data inside the current one.

Any clues are welcomed or could you please tell me where I can get some direct help?

Thanks in advance.

[douglasmaltby]

Many thanks to Thorsten! He pointed me to the troubleshooting section of the SP04 getting started guide, which helped resolve the "gateway timeout" issue in my SDH 1.4 trial.

• https://caldocs.hana.ondemand.com/caldocs/help/Getting_Started_Data_Hub_SP04.pdf on page 30

In my case, there was NO difference in the kubectl APP_SELECTOR value and what was already in the (load balancer) internallb.yaml file, but simply re-applying the exact same config (as root) with command “kubectl -n datahub apply -f internallb.yaml” seemed to resolve the issue!

• There was a warning that "Warning: kubectl apply should be used on resource created by either kubectl create --save-config or kubectl apply"
• But it seemed to apply the "change": service "data-hub-flow-agent" configured

Also note that there are some "extra" steps in the gateway troubleshooting section that don't apply (from section 4 “Connecting to Google Cloud Storage”). Thorsten has fixed the source document and it'll be updated online soon.

After re-applying the unchanged internallb.yaml, and checking the Data Hub Cockpit again, it worked! No "Gateway Timeout". The System Status section was refreshed, and I could see in Landscape Management that the VORA_SYS_DEFAULT connection was working, and Discovery > Profiled data has all my previously profiled data from Vora, GCS and AWS available (stored in a GCS bucket, not HDFS).

With both values the same, it doesn't seem like re-applying the .yaml would have worked, but it did...

It's possible that the root cause was something transient in Google's networking, but before I did this, the SDH Cockpit got the Gateway timeout error, and afterwards it worked perfectly again.

Thanks again, Thorsten!

Doug

[schneidertho]

Hi Haley,

sorry for keeping you waiting so long. Missed your question. Is this still occurring? I don't have an immediate idea.

The HANA connection is a direct connection, so flowagent should not be the root cause. To somehow find out what is going wrong, I would probably need some more details about the connections (and network setup).

Please ping me if you still need help, either via this question or direct message.

Cheers
Thorsten

[douglasmaltby]

Hi Thorsten,

I have the exact same issue Haley illustrates above, but with the 1.4 trial. A few days ago, everything was fine, but now this gateway issue.

I haven't been able to resolve it, but have captured what I think might be the relevant network, firewall, routes and kubernetes service endpoint details, which might not be suitable to share here.

I also reviewed the OSS release restriction note on 1.4: https://launchpad.support.sap.com/#/notes/2649485 and the prior 1.3 note: https://launchpad.support.sap.com/#/notes/2602716. Only the 1.3 note describes the issue, but not 1.4. Notably, though, the comment with "gateway" says:

• Not possible to remove systems in SAP Data Hub cockpit which are in diconnected state. Removal of systems from the landscape management in SAP Data Hub Cockpit may fail with a Gateway Timeout error in case the remote SAP Data Hub SDI Adapter is not reachable. Ensure that the Adapter is running before removing the system.

I'll try to email or DM you (followed on Twitter) privately and share the situation details this weekend.

Thanks,

Doug

[yu_chen_10]

@Halley, I have used the Trial System in Cloud for some time. It was okay. But now I have the same problem with "Gateway timeout". Have you already got the solution? Thanks.

[haley_zhang]

Haven't checked that for some time since the question was raised.

It might become fine after some days without any solution, or it might not. I haven't completely figure it out so far...