I have 10 nodes and I use; CephFS, RBD and RGW clients and all of my clients are 14.2.16 Nautilus. My clients, MONs, OSDs are on the same servers. I have constant usage: 50-300MiB/s rd, 15-30k op/s rd --- 100-300MiB/s wr, 1-4 op/s wr. With the allocator issue it's highly possible to get slow ops and OSD down while upgrading. I will start synthetic load while upgrading the test environment to be sure. Thanks for the tips btw. Stefan Kooman <stefan@xxxxxx>, 10 Kas 2021 Çar, 19:36 tarihinde şunu yazdı: > On 11/10/21 17:22, mhnx wrote: > > Hello Igor. Thanks for the answer. > > > > There are so many changes to read and test for me but I will plan an > > upgrade to Octopus when I'm available. > > > > Is there any problem upgrading from 14.2.16 ---> 15.2.15 ? > > I have upgraded a few test clusters from 14.2.16 to 15.2.15. I hit no > issues. But those have barely any load. Some things you might want to > check beforehand: > > Do any default settings change from 14.2.16 tot 15.2.15 (i.e. > bluefs_buffered_io=true) > > Not sure what clients you have, but in order to be able to mitigate > "AUTH_INSECURE_GLOBAL_ID_RECLAIM" you ideally want to have all your > clients upgraded before you upgrade your Ceph cluster. Otherwise you > will have to mute the warning you will get. > > We have the following set: > > advanced osd_fast_shutdown=false > > We noticed if that was set to true we would get slow ops when rebooting > OSD storage nodes (mimic -> Nautilus). Some users have hit slow ops > issues when upgrading from Nautilus -> Octopus. See this thread [1] > (still unresolved). > > The following settings should work: > > osd_fast_shutdown = true > osd_fast_shutdown_notify_mon = false > > But apparently not in all circumstances. > > YMMV. > > Gr. Stefan > > [1]: > > https://lists.ceph.io/hyperkitty/list/ceph-users@xxxxxxx/thread/J6CRWRBIVZKEVWOTC2MIA2A4E5PHJ6SE/ > _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |