Pavel, it's not about the stack trace, it's about the right exception to be thrown to the user. JCache specifies that if any exception is thrown by the EntryProcessor, this exception should be wrapped into EntryProcessorException. So, we need to inform the client that the error was caused by EntryProcessor. First option to do this: introduce new error code, second option: use dedicated "success" flag.
ср, 8 мая 2024 г. в 07:35, Pavel Tupitsyn <ptupit...@apache.org>: > > > In case of such response we need some new > > error code to distinguish entry processor exceptions and general > client/server exceptions. > > I don't think we do this anywhere else (e.g. Compute). > Full stack trace should be present in the server logs, or sent back to the > client when sendServerExceptionStackTraceToClient is enabled, so we have > all the means to find the root cause, right? > > On Tue, May 7, 2024 at 9:32 PM Alex Plehanov <plehanov.a...@gmail.com> > wrote: > > > Hi Igniters! > > > > I've prepared a proposal about implementation of invoke/invokeAll > > cache operations for thin clients. All proposed changes to the > > protocol and client API (currently for java thin client only) are > > described in IEP-122 [1] > > > > Please have a look and share your thoughts. > > > > [1] > > https://cwiki.apache.org/confluence/display/IGNITE/IEP-122+Thin+client+invoke%2C+invokeAll+cache+operations > >
