Hi Shlshya, Do you see the warnings with enabled 1st level cache? Could you provide a test case to let me reproduce the case?
Best regards, Artem Orobets * Orient Technologiesthe Company behind OrientDB* 2014-03-12 8:57 GMT+02:00 Shishya <[email protected]>: > Hi, > > I see such warnings in 1.7rc1 as well. > Does it impact any data integrity? > > Regards > > > On Tuesday, September 17, 2013 10:25:49 PM UTC+5:30, [email protected]: >> >> We adjusted the test code to explicitly turn on the level 1 cache and the >> warning is still shown. >> >> We tried again in the main code and now we're not seeing any difference >> (no additional exceptions). We did do a bit of refactoring and also updated >> again to the latest SNAPSHOT (dated this morning), so that may have changed >> the behavior. So now, however, we are seeing no difference in output with >> level 1 cache enabled or disabled. The warning is shown either way. >> >> On Tuesday, September 17, 2013 9:35:14 AM UTC-5, Andrey Lomakin wrote: >>> >>> Hi Mike, >>> If you disable 1-st level cache you will have these exception because it >>> is main responsibility of given cache to prevent such cases. >>> >>> You wrote that you got exceptions when you enbable 1-st level cache >>> could you send them ? >>> >>> >>> On Tue, Sep 17, 2013 at 12:04 AM, <[email protected]> wrote: >>> >>>> We've created code that reproduces this issue. I've created a Gist for >>>> it at https://gist.github.com/mikeosterlie/f2c557aadb54370623da. When >>>> the transaction is committed the warning is shown. >>>> >>>> >>>> On Monday, September 16, 2013 3:37:54 PM UTC-5, [email protected]: >>>>> >>>>> I think we've got this narrowed down. We're going to try to replicate >>>>> the issue in a testcase, but the flow looks like this: >>>>> >>>>> We have a tree (doesn't matter, but it's an AVL tree implementation) >>>>> made of vertices and edges. As part of a transaction we are deleting an >>>>> edge between two nodes and creating an edge between a different set of >>>>> nodes. We are using plocal so that RIDs are not reused, but the newly >>>>> created edge is getting the same RID has the just deleted edge. This >>>>> produces that warning as we now have two edges in the transaction with the >>>>> same RID, but are different instances. >>>>> >>>>> I thought plocal did not reuse RIDs. >>>>> >>>>> On Monday, September 16, 2013 10:45:46 AM UTC-5, Andrey Lomakin wrote: >>>>>> >>>>>> >>>>>> On Mon, Sep 16, 2013 at 6:18 PM, <[email protected]> wrote: >>>>>> >>>>>>> The warning was still displayed with level 1 cache on, and we ran >>>>>>> into a lot of varied Exceptions as well. That might point to some logic >>>>>>> issues in our code, and we have identified a potential cause based on >>>>>>> Sylvain's posts as our logic could load the same record into two >>>>>>> different >>>>>>> variables. >>>>>> >>>>>> >>>>>> I am looking forward for your updates after code verification. >>>>>> It would be good to fix this issue till 1.6 will be released. >>>>>> >>>>>> -- >>>>>> Best regards, >>>>>> Andrey Lomakin. >>>>>> >>>>>> Orient Technologies >>>>>> the Company behind OrientDB >>>>>> >>>>>> -- >>>> >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "OrientDB" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> For more options, visit https://groups.google.com/groups/opt_out. >>>> >>> >>> >>> >>> -- >>> Best regards, >>> Andrey Lomakin. >>> >>> Orient Technologies >>> the Company behind OrientDB >>> >>> -- > > --- > You received this message because you are subscribed to the Google Groups > "OrientDB" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
