UP. Problem still exist... :( środa, 24 lutego 2021 o 18:15:11 UTC+1 [email protected] napisał(a):
> Does anyone have any experience to share regarding DSpace 7? We are really > suffering with DSpace 6: > > - Discovery indexing of 90,000 items in DSpace 6.3 takes four or five > hours, where it used to take one or two in Dspace 5.8 on the same hardware > - Harvesting all our items via the REST API /items endpoint in DSpace 6.3 > takes ten hours, where it used to take half of that in DSpace 5.8 (though > the DSpace REST API has never been performant) > - Ingesting a CSV file with 1,000 metadata-only items takes one hour in > DSpace 6.3 > > *At this rate I'm wondering who I can pay to troubleshoot the Hibernate > issues in DSpace 6*. Does DSpace 7 still have these issues? It's starting > to be a real drag for us. I don't know how owners of mid-to-large DSpace > repositories survive on the modern web with performance profiles like this. > I will investigate nginx caching of API responses to try to lessen the load. > > Regards, > > On Wed, Jan 27, 2021 at 2:27 PM Karol Sokalski <[email protected]> > wrote: > >> Alan, >> >> thank You, for sad, but helpfully answer. Maybe DSpace7 will be without >> this problem... >> Cheers! >> >> Karol >> wtorek, 26 stycznia 2021 o 20:57:57 UTC+1 [email protected] napisał(a): >> >>> Dear Karol, >>> >>> Yes I noticed the same thing when moving from DSpace 5 to 6 in 2020: >>> indexing performance, REST API responses, item submission, etc are all much >>> slower in DSpace 6. More specifically, indexing is three times slower in >>> our repository of ~90,000 items, harvesting the entire repository via REST >>> API takes ~10 hours, and editors have complained about submission times. >>> Etc etc. >>> >>> In a February, 2020 thread on this list¹ I did an extensive performance >>> analysis between DSpace 5.10 and 6.4 Discovery indexing to try to >>> understand the difference. I'm attaching the flame graphs from that >>> research here (download them and open in Firefox, they are interactive >>> SVGs). The sample time was sixty seconds. The length of a bar is the time >>> the function spent on the CPU. The height grows with the spawning of >>> children. This clearly shows that Hibernate has a significant overhead in >>> DSpace. I am not a Java developer (let alone an expert in Hibernate). I >>> also hope that someone can figure this out. >>> >>> Cheers, >>> >>> ¹ See my thread with subject "Slow Discovery indexing performance in >>> DSpace 6" >>> >>> On Tue, Jan 26, 2021 at 12:11 PM Karol Sokalski <[email protected]> >>> wrote: >>> >>>> Hi, >>>> >>>> two weeks ago i upgraded DSpace from 5.4 to 6.3. Everything looks >>>> pretty good, but today, webadmin told me "creating new collections takes a >>>> long time, or is imposible (Internel server error) when i push "next"" >>>> >>>> [image: dspaceCollectionDescription.png] >>>> I enabled debug mode, and i saw that, everytime when i push "Next" in >>>> dspace.log starting a lot of operations like this connected with hibernate: >>>> >>>> 2021-01-25 17:50:26,746 DEBUG >>>> *org.hibernate.engine.internal.Collections* @ Collection found: [ >>>> *org.dspace.eperson.Group*.*groups*#4491bccf-303c-434343395-823497b-d8243234092d655fb3], >>>> >>>> was: >>>> [org.dspace.eperson.Group.groups#4491bccf-303c-43434395-897b-d802343292d655fb3] >>>> >>>> (uninitialized) >>>> 2021-01-25 17:50:26,746 DEBUG org.hibernate.engine.internal.Collections >>>> @ Collection found: >>>> [*org.dspace.eperson.Group.parentGroups*#4491bccf-303c-439234325-89234427b-d8092d655fb3], >>>> >>>> was: >>>> [org.dspace.eperson.Group.parentGroups#4491bccf-303c-43924325-8923427b-d8092d655fb3] >>>> >>>> (uninitialized) >>>> 2021-01-25 17:50:26,746 DEBUG org.hibernate.engine.internal.Collections >>>> @ Collection found: >>>> [org.dspace.eperson.Group.supervisedItems#4491bccf-303c-439234325-897b-d8092d655fb3], >>>> >>>> was: >>>> [org.dspace.eperson.Group.supervisedItems#4491bccf-303c-43243295-897b-d8092d655fb3] >>>> >>>> (uninitialized) >>>> 2021-01-25 17:50:26,746 DEBUG org.hibernate.engine.internal.Collections >>>> @ Collection found: >>>> [org.dspace.content.DSpaceObject.handles#0d2abd2f-e192-4a8234322-9e49-11e2e0f05c18], >>>> >>>> was: >>>> [org.dspace.content.DSpaceObject.handles#0d2abd2f-e192-4a8234322-9e49-11e2e0f05c18] >>>> >>>> (uninitialized) >>>> 2021-01-25 17:50:26,746 DEBUG org.hibernate.engine.internal.Collections >>>> @ Collection found: >>>> [org.dspace.content.DSpaceObject.metadata#0d2abd2f-e192-4a243282-9e49-11e2e0f05c18], >>>> >>>> was: >>>> [org.dspace.content.DSpaceObject.metadata#0d2abd2f-e192-4a2342382-9e49-11e2e0f05c18] >>>> >>>> (uninitialized) >>>> >>>> The problem exist with: >>>> - Create collection when i check >>>> * The submission workflow will include an *accept/reject* step >>>> * The submission workflow will include an *accept/reject/edit >>>> metadata* step >>>> * *The submission workflow will include an edit metadata step* >>>> * **This collection will have delegated collection administrators >>>> The same, when i edditing new collection first time >>>> (accept/reject/edit etc). >>>> >>>> It looks like problem with new mechanism "hibernate". Is posible to >>>> increase performance ? For example by modify (hibernate.cfg.xml) or >>>> something else? >>>> >>>> Thanks, >>>> >>>> Karol >>>> >>>> >>>> Thanks, >>>> >>>> -- >>>> All messages to this mailing list should adhere to the DuraSpace Code >>>> of Conduct: https://duraspace.org/about/policies/code-of-conduct/ >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "DSpace Technical Support" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/dspace-tech/e3a93319-958e-41fa-a447-ddcf5cc8af5dn%40googlegroups.com >>>> >>>> <https://groups.google.com/d/msgid/dspace-tech/e3a93319-958e-41fa-a447-ddcf5cc8af5dn%40googlegroups.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>> >>> >>> -- >>> Alan Orth >>> [email protected] >>> https://picturingjordan.com >>> https://englishbulgaria.net >>> https://mjanja.ch >>> >> -- >> All messages to this mailing list should adhere to the DuraSpace Code of >> Conduct: https://duraspace.org/about/policies/code-of-conduct/ >> --- >> You received this message because you are subscribed to the Google Groups >> "DSpace Technical Support" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> > To view this discussion on the web visit >> https://groups.google.com/d/msgid/dspace-tech/c9ad3191-c1b8-457c-8c1b-ac0b34aac3fdn%40googlegroups.com >> >> <https://groups.google.com/d/msgid/dspace-tech/c9ad3191-c1b8-457c-8c1b-ac0b34aac3fdn%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > > > -- > Alan Orth > [email protected] > https://picturingjordan.com > https://englishbulgaria.net > https://mjanja.ch > -- All messages to this mailing list should adhere to the Code of Conduct: https://duraspace.org/about/policies/code-of-conduct/ --- You received this message because you are subscribed to the Google Groups "DSpace Technical Support" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/dspace-tech/ba8d7bc2-6b4c-4dba-a5c1-d426cb69e41en%40googlegroups.com.
