Thank you, We have tried the way in the thread, it did not work, because the jcache contained in geronimo-jcache_1.0_spec-1.0-alpha-1.jar has the different serivalVersionUID in yarn library, which we cannot simply replace it. The jcache can have different version in the cluster, we have to simply upload geronimo-jcache_1.0_spec-1.0-alpha-1.jar in order to overwrite the yarn’s one.
From: Вячеслав Коптилин <[email protected]> Date: Wednesday, May 24, 2023 at 11:55 AM To: [email protected] <[email protected]> Subject: Re: incompatible class error Hello, In general, you need to provide the right version of JCache API on all nodes. Please take a look at https://issues.apache.org/jira/browse/IGNITE-11986 (The issue mentioned in the ticket looks very similar to your case) Thanks, S. пн, 15 мая 2023 г. в 21:00, Jiang Jacky <[email protected]<mailto:[email protected]>>: Hello, can someone help me about this? Thank you. From: Jiang Jacky <[email protected]<mailto:[email protected]>> Date: Saturday, May 13, 2023 at 1:58 AM To: [email protected]<mailto:[email protected]> <[email protected]<mailto:[email protected]>> Subject: incompatible class error Hello, I am running Gridgain Ignite with spark, I am using spark 2.4.5 and Hadoop 3.1.1 version. My gridgain cluster is 2.7.0 version However, I am stuck at below error: Caused by: java.io.InvalidClassException: javax.cache.configuration.MutableConfiguration; local class incompatible: stream classdesc serialVersionUID = 201405, local class serialVersionUID = 201306200821 After investigation, I find it is caused by the jar from yarn lib geronimo-jcache_1.0_spec-1.0-alpha-1.jar I cannot delete the jar, because it is part of YARN cluster. So, anyone has the experience about how to fix the issue, please let me know. Thank you. Jacky
