well, i guess this was not a simple a question as i thought. i need to do some testing.
/mrg On Feb 14, 2014, at 4:55 PM, Rich Megginson <[email protected]> wrote: > On 02/14/2014 02:52 PM, Michael Gettes wrote: >> i guess i wasn’t clear in my question - i am not seeing a problem with paged >> results. > > Sorry - I meant - are you seeing a problem currently with uid=* searches, > without paged results? If you are seeing high memory usage - does the memory > usage go down when the search is complete? Is this memory that is not in a db > or entry cache? > >> and i think you answered my question. with paged results one should be able >> to perform a query with large result sets and not have large memory >> concerns, right? > > I am not sure. That is, paged results doesn't necessarily help with the > memory usage. > >> this would be the answer i am looking for. >> >> /mrg >> >> On Feb 14, 2014, at 4:13 PM, Rich Megginson <[email protected]> wrote: >> >>> On 02/14/2014 02:04 PM, Michael Gettes wrote: >>>> I did some searching to try and answer the following question and the >>>> answer is not apparent to me. >>>> >>>> I have a directory with hundreds of thousands of entries. I have an app >>>> which, due to its behavior, needs to >>>> be able to search for very large sets of users but I don’t want it to be >>>> able to get ALL users and therefore exhaust >>>> all memory on the ldap server. >>>> >>>> If the app performs a search on uid=* and the it’s a paged search, does >>>> this have a positive impact on how >>>> memory is used to handle the results so it doesn’t exhaust all memory? >>>> >>>> I’m trying to keep the scenario simple so I hope you appreciate the >>>> question I am getting at. >>> The primary reason for using paged results on the server side is to allow >>> you to perform searches which otherwise would have hit the lookthrough >>> limit, and caused high CPU spikes. It should also reduce memory usage, >>> although I'm not sure why it should exhaust memory in the first place. Are >>> you seeing the directory server run out of memory with these types of >>> searches? >>> >>>> thanks! >>>> >>>> /mrg >>>> >>>> -- >>>> 389 users mailing list >>>> [email protected] >>>> https://admin.fedoraproject.org/mailman/listinfo/389-users >>> -- >>> 389 users mailing list >>> [email protected] >>> https://admin.fedoraproject.org/mailman/listinfo/389-users >> -- >> 389 users mailing list >> [email protected] >> https://admin.fedoraproject.org/mailman/listinfo/389-users > > -- > 389 users mailing list > [email protected] > https://admin.fedoraproject.org/mailman/listinfo/389-users -- 389 users mailing list [email protected] https://admin.fedoraproject.org/mailman/listinfo/389-users
