This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

large Solr8 index trips up and now Solr7 is at same % done. just switch?

Solr8 initial index run got to 81% and then something set it and Solr7 back, first seeing both at 79%

Some how the Solr8 run also caused Veeam to struggle with its backup suddenly up to 48 hours after the first day or so. We've paused that backup for now

they are now both in lock step, up to 84% after just over a day.

Options I am debating and would like your thoughts

- just making the switch to Solr8 and leave Solr7 in the dust?

- add cores,  only two on it.   of course this requires downing the server for a bit.

- just rebooting the vm after a problem shut down.

________________________

Andy of KonecnyConsulting.ca in Toronto
Please use the "Like" and/or "Verified Answers" as appropriate as that helps us all.

  • 0  

    Yes, switch to Solr8!

    I would add two more cores - but maybe later (It depends a little bit on environment size).

    I know you have been part of this discussion but do not forget ...  Performance difference SOAP with and w/o SSL 


    Use "Verified Answers" if your problem/issue has been solved!

  • 0   in reply to   

    We have cores(and RAM) to spare. We can run all the guests on one host if needed, with the 10GB LAN & SAN links becoming the main bottleneck

    SSL not an issue.  Stopped at 18.2 just before having to force it.   Their clients and staff forcing migration to M365 and all that fun I am preparing for.

    ________________________

    Andy of KonecnyConsulting.ca in Toronto
    Please use the "Like" and/or "Verified Answers" as appropriate as that helps us all.

  • 0   in reply to   

    See this Retain TID titled "Improving Indexing Performance By Adding Threads And Cores" at 

    https://support.microfocus.com/kb/doc.php?id=7019817 

    Disk IOPS on the .../index8/ file system is a big factor.

  • 0   in reply to   

    I think I will be doing that manual pump of index threads.   This system isn't touched that often that 2 cores was enough for day to day usage, and don't really want to give it too much more if I can.  Perhaps with the 4 cores, 8 threads for indexing? 

    we were so close to finishing, was at 97% this morning,  now I see both are back to 78%

    <sigh>   deeper dive tomorrow as other things beckon

    This SAN auto sets what blocks migrate to spinning, otherwise are on SSD.  Will be watching the IOPs

    ________________________

    Andy of KonecnyConsulting.ca in Toronto
    Please use the "Like" and/or "Verified Answers" as appropriate as that helps us all.

  • Suggested Answer

    0   in reply to   

    Finally fixed with help from support.  Not sure which exactly fixed things, but the path is worth knowing if others hit his issue.

    - increasing cores and threads as discussed above

    - Drop Solr logging from debug to 'just' error

    - Switched to Solr8, made a backup copy of the Solr7 index, and then removed the Solr7 indexing.

    Those made the indexing run so much faster, such that it avoided what ever was tripping up, possibly relating to both Indexers running, or just time.

    One last observation also led to an Idea

    So after all the tricks I learned in this, the next upgrade/migration I do will likely just go smoothly without needing this. <sigh>

    ________________________

    Andy of KonecnyConsulting.ca in Toronto
    Please use the "Like" and/or "Verified Answers" as appropriate as that helps us all.

  • 0   in reply to   

    Thanks for the update of what did it for you.   

    Do update us on your next upgrade you do to verify if it does go smoothly from the start.