Hello! Regardless of extension your thread dumps almost do not contain anything relevant.
There is one guess that can be made: 0x000007fa7fe069c1 ignite_dll!?save_object_ptr@ ?$pointer_oserializer@Vportable_binary_oarchive@@VVolatilityCubeModel@credit @@@detail@archive@boost@@EEBAXAEAVbasic_oarchive@234@PEBX@Z + 0x4c901 Maybe it's something about taking a lot of time to serialize your functions. Still it's hard to say without Java thread dump. Regards, -- Ilya Kasnacheev пн, 25 февр. 2019 г. в 11:39, F.D. <[email protected]>: > Hi, > in my previous message there was the dump of the 4 servers. Maybe the > extension of the files is misleading. > > Thanks, > F.D. > > On Mon, Feb 25, 2019 at 7:48 AM Ilya Kasnacheev <[email protected]> > wrote: > >> Hello! >> >> Please post Java stack traces collected with jstack utility. Otherwise >> there's nothing there for us to see, unfortunately. >> >> Regards, >> -- >> Ilya Kasnacheev >> >> >> чт, 21 февр. 2019 г. в 18:33, F.D. <[email protected]>: >> >>> Hi, >>> I reduced the grid to only 4 server, so it was easier to collect dumps. >>> In the logs you'll find 3 attempts, the first and the third are dump of the >>> problem, the second it's a dump with the server in idle. >>> >>> Thanks for your support! >>> >>> On Thu, Feb 21, 2019 at 10:58 AM Ilya Kasnacheev < >>> [email protected]> wrote: >>> >>>> Hello! >>>> >>>> Preferably all nodes in the cluster. >>>> >>>> Regards, >>>> -- >>>> Ilya Kasnacheev >>>> >>>> >>>> чт, 21 февр. 2019 г. в 12:39, F.D. <[email protected]>: >>>> >>>>> Hi, >>>>> >>>>> thread log of the client? >>>>> >>>>> On Thu, Feb 21, 2019 at 10:29 AM Ilya Kasnacheev < >>>>> [email protected]> wrote: >>>>> >>>>>> Hello! >>>>>> >>>>>> Can you collect some thread dumps (Java ones, using jstack) during >>>>>> those 10 seconds freezes? >>>>>> >>>>>> Regards, >>>>>> -- >>>>>> Ilya Kasnacheev >>>>>> >>>>>> >>>>>> ср, 20 февр. 2019 г. в 20:18, F.D. <[email protected]>: >>>>>> >>>>>>> Hi igniters, >>>>>>> I've a problem when I launch a sequence of about 350 compute >>>>>>> funcions in my cluster of 10 node. I register a delay in the first >>>>>>> launcher >>>>>>> od about 10 sec. after that, it works smoothly. >>>>>>> If I launch the same numbers of task on a single node, there 's no >>>>>>> delay. >>>>>>> It seems like a problem in the discovery process of the whole grid. >>>>>>> >>>>>>> Do you have any ideas? >>>>>>> >>>>>>> Thanks, >>>>>>> F. D. >>>>>>> >>>>>>
