AFM supports data migration between the two different file systems from the same cluster using NSD protocol. AFM based migration using the NSD protocol is usually performed by remote mounting the old filesystem (if not in the same cluster) at the new cluster's gateway node(s). Only gateway node is required to mount the remote filesystem.
Some recent improvements to the AFM prefetch 1. Directory level prefetch, users no longer required to provide list files. Directory prefetch automatically detects the changed or new files and queues only the changed files for the migration. Prefetch queuing starts immediately, and does not wait for the full list file/directory processing unlike in the earlier releases (pre 5.0.2). 2. Multiple prefetches for the same fileset from different gateway nodes. (will be available in 5.0.3.x, 5.0.2.x). User can select any gateway node to run the prefetch for a fileset, or split list of files or directories and execute them from the multiple gateway nodes simultaneously. This method gets good migration performance and better utilization of network bandwidth as the multiple streams are used for the transfer. 3. Better prefetch queueing statistics than previous releases, provides total number of files, how many queued, total amount of data etc.. ~Venkat ([email protected]) From: "Lyle Gayne" <[email protected]> To: gpfsug main discussion list <[email protected]> Date: 04/01/2019 07:35 PM Subject: Re: [gpfsug-discuss] A net new cluster Sent by: [email protected] Yes, native GPFS access can be used by AFM, but only for shorter distances (10s of miles, e.g.). For intercontinental or cross-US distances, the latency would be too high for that protocol so NFS would be recommended. Lyle "Marc A Kaplan" ---03/29/2019 03:05:53 PM---I don't know the particulars of the case in question, nor much about ESS rules... From: "Marc A Kaplan" <[email protected]> To: gpfsug main discussion list <[email protected]> Date: 03/29/2019 03:05 PM Subject: Re: [gpfsug-discuss] A net new cluster Sent by: [email protected] I don't know the particulars of the case in question, nor much about ESS rules... But for a vanilla Spectrum Scale cluster -. 1) There is nothing wrong or ill-advised about upgrading software and then creating a new version 5.x file system... keeping any older file systems in place. 2) I thought AFM was improved years ago to support GPFS native access -- need not go through NFS stack...? Whereas your wrote: ... nor is it advisable to try to create a new pool or filesystem in same cluster and then migrate (partially because migrating between filesystems within a cluster with afm would require going through nfs stack afaik) ... _______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at spectrumscale.org http://gpfsug.org/mailman/listinfo/gpfsug-discuss _______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at spectrumscale.org https://urldefense.proofpoint.com/v2/url?u=http-3A__gpfsug.org_mailman_listinfo_gpfsug-2Ddiscuss&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=92LOlNh2yLzrrGTDA7HnfF8LFr55zGxghLZtvZcZD7A&m=rdsJfQ2D_ev0wHZkn4J-X3gFEMwJzwKuuP0EVdOqShA&s=4Du5XtaI8UBQwYJ-I772xbA5kidqKoJC-XasFXwEdsM&e=
_______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at spectrumscale.org http://gpfsug.org/mailman/listinfo/gpfsug-discuss
