2024-09-05T03:59:09.678Z

<Milind Changire> is there an OSD op that could allocate disk blocks without actually writing to the backing objects ?
this is w.r.t. optimally implementing `fallocate` for CephFS

2024-09-05T15:16:25.755Z

<Kaleb> @Ken Dreyer there was some post somewhere from you about a  jira ticket where you wanted me to confirm the inclusion of ganesha gmonitoring rpm in upstream containers? Would you repost that?  Thanks

2024-09-05T17:02:33.098Z

<Matt Benjamin> seems like I should not see this (ceph/main) after successful cmake run:

2024-09-05T17:02:37.777Z

<Matt Benjamin> > -- Build files have been written to: /lv2tb/ceph-cp/build
> [mbenjamin@fedora build]$ ninja
> ninja: error: '/lv2tb/ceph-cp/src/nvmeof/gateway/control/proto/gateway.proto', needed by '/lv2tb/ceph-cp/build/src/gateway.pb.cc', missing and no known rule to make it

2024-09-05T17:03:06.717Z

<Matt Benjamin> I have head from team members that they get around this by skipping nvmeof

2024-09-05T17:08:28.507Z

<Matt Benjamin> if we're not going to fix that, could we have WITH_NVMEOF_GATEWAY_MONITOR_CLIENT default to NO/OFF?

2024-09-05T18:37:25.927Z

<avasiliev> I used it as main allocator in our project. Worked great but I cannot tell how it would compare with others. Probably won't be drastically different.

It has some nice features like custom heaps that don't share locks with the main heap and also implements allocator concept that is neat. But we used some task specific ones instead. 

2024-09-05T21:18:39.507Z

<Ken Dreyer> Sure, <https://github.com/ceph/ceph-container/issues/2235>