Hello! As far as my understanding goes, this is how readThrough work. It may update only subset of REPLICATED cache partitions across cluster.
Regards, -- Ilya Kasnacheev пт, 28 февр. 2020 г. в 21:09, Prasad Bhalerao <[email protected] >: > Hi, > > We have not set any expiration/ eviction policy. > We were getting false for key "key1", so we checked the key/value by > querying cache using web console and the value is present in cache. But > still getting false for key "key1" on subsequent containskey execution. > > But for key "key2" we were getting true. > > Thanks, > Prasad > > On Fri 28 Feb, 2020, 10:58 PM Denis Magda <[email protected] wrote: > >> Hi Akash, >> >> Do you execute the cache.contains() method after reading-through the >> record with cache.get()? Do you have any expiration/eviction policies set >> that may purge the record from memory after being loaded from disk? >> >> - >> Denis >> >> >> On Fri, Feb 28, 2020 at 9:11 AM Akash Shinde <[email protected]> >> wrote: >> >>> Hi, >>> I am using Ignite 2.6 version. >>> >>> I have partitioned cache, read-through and write-through is enabled. >>> Back-up count is 1 and total number of server nodes in cluster are 3. >>> >>> When I try to get the data from a cache for a key using cache.get(key) >>> method, ignite reads the value from database using provided cache loader >>> and returns the value by read-through approach. >>> >>> But when I execute cache().containsKey(key) on client node, I get false. >>> >>> But the strange this is this behavior is not same for all keys of the >>> same cache. >>> For key1 I get false but for key2 I get true. But both the keys are >>> present in cache. >>> >>> I executed the SQL on every node (one node at a time) using web-console, >>> I got data present only on one node out of three. This seems to be primary >>> for this particular key. >>> >>> Can someone please advise why this is happening? Is it a bug in ignite? >>> This seems to be a very basic case. >>> >>> >>> >>> Thanks, >>> Akash >>> >>
