Thanks Shawn for the information. Is it possible that the segment merging which happens during hardcommit can open a new searcher thereby making index updates visible during search?
On 2022/05/26 17:19:21 Shawn Heisey wrote: > On 5/26/22 10:14, Poorna Murali wrote: > > In our configuration, we are not using soft commit at all and autocommit is > > done by hard commit with opensearcher = false which will not create a new > > searcher. > > If you never do a commit that opens a new searcher, then you will never > see the result of adds, updates, or deletes in your searches. If you > are making changes to your index and those changes ARE visible to > clients, then SOMETHING is opening a new searcher, which is only going > to happen when there's a commit. > > > So, I am not able to understand how is the fieldcache entries are getting > > cleared. Is there any other way that this can happen as in when segment > > merging happens during hard commit? > > The FieldCache is not built or maintained by Solr. It is a Lucene > construct. And I think in current versions, it is actually no longer > named FieldCache, but I do not remember the new name. > > Maybe somebody here knows enough about it to tell you what might be > happening. > > My guess is that there are explicit manual commits happening. That might > be part of an update request or a separate API call. > > Thanks, > Shawn > >
