Hi Huijun, On our side, we are definitely able to saturate the link bandwidth with SSDs. I actually tested a similar configuration last Friday and it worked well. Maybe you should try with a single SS

Hi there, Please note that the agenda for the 4th annual DAOS User Group meeting is now available online: https://wiki.hpdd.intel.com/display/DC/DUG20 I am very excited by the diversity and number of

Hi Anton, The DAOS control plane does not support NVMe-oF target yet, that’s why you don’t see them via storage scan. That being said, my hope was that, by changing the daos_nvme.conf manually as you

FYI. If you need any help, feel free to ping us 😊 Subject: IO500 SC20 Call for Submission Call for IO500 Submission Deadline: 30 October 2020 AoE Stabilization period: 1st October -- 9th October 2020

Really sorry for the spam, but I used the wrong timezone when creating the invite ☹ The last one that was sent to the mailing list (attached for reference) is the correct one. Cheers, Johann From: <da

Hi there, Please find below the Webex meeting invite for the DUG’20. The agenda will be posted there: https://wiki.hpdd.intel.com/display/DC/DUG20 Cheers, Johann From: <daos@daos.groups.io> on behalf

Glad it helped. You might also want to remove /tmp/daos_server.log which is probably pretty big. As for getting even better performance, I would increase the block size to write more and maybe try wit

Could you please change “log_mask: DEBUG” to “log_mask: ERR”, restart the server and rerun ior? Cheers, Johann From: <daos@daos.groups.io> on behalf of "anton.brekhov@..." <anton.brekhov@rsc-t

Hi Anton, Saw your results on the other email. A few comments: We indeed do expect DFS to perform better than dfuse. Could you please try to run some more tests with different tsize and process counts

Anton, aren’t the header files in the *-devel RPMs as mentioned by Steffen? Cheers, Johann From: <daos@daos.groups.io> on behalf of "anton.brekhov@..." <anton.brekhov@...> Reply-To: "d

Thanks Mohamad & Steffen. Can we easily change the IOR build to assume “--with-daos” = “--with-daos=/usr --with-cart=/usr”? TBH, I would also just assume that CART and DAOS libs are located at the sam

I was hoping that --with-daos would be enough if everything is installed at the default location via RPMs. @Chaarawi, Mohamad any thoughts? Cheers, johann From: <daos@daos.groups.io> on behalf of "ant

Ok, glad that it works now for dfuse. Cheers, Johann From: <daos@daos.groups.io> on behalf of "anton.brekhov@..." <anton.brekhov@...> Reply-To: "daos@daos.groups.io" <daos@....

Hey Anton, I assume that you have built ior with the DFS backend support as specified in https://github.com/hpc/ior/blob/master/README_DAOS, right? Although I think it should work, I have never tried

Hi Anton, Are you able to create files manually under the dfuse mountpoint? Does ior work with file per process (-F)? Cheers, Johann From: <daos@daos.groups.io> on behalf of "anton.brekhov@...

Hi there, Please find attached a small test that you can compile as follows: $ gcc -o test -L${DAOS_PATH}/lib64 -I${DAOS_PATH}/include -ldaos -lgurt -ldaos_common -luuid ./test.c And then run: $ D_LOG

As pointed out by Niu, the main gap is really on the DAOS control plane (i.e. dmg storage scan & support in the yaml config file). SPDK has built-in support for a NVMe-oF initiator/host (see https://s

Hi Anton, Yes, I think it is fine. You probably leaked pools from prior runs that show up in pool listing while the test is assuming that no other pools are present. I think you are good to go. Cheers

It sounds like the DAOS server is not able to register memory to initiate a RDMA. Could you please tell me more about the network and storage you use on the server? Optane PMEM or DRAM? Also, what ver

Looks better. The network stack seems to be properly initialized now. The failure you see now (“undefined symbol: dmg_pool_create”) is related to the C wrapper of the dmg command which is used to crea