Christian, Like any AFM deployment in Spectrum Scale, you define how much local working space you want to allocate in the Spectrum Scale filesystem and files will be cached on demand into this 'working set' and will never exceed this upper bound.
If you create a new file on this filesystem (in this fileset) then it will automatically be queued to be replicated onto the backend. If you make continuous changes to a file then setting afmAsyncDelay can reduce the number of writes to the backend object store. The local copy though will be kept and only stubbed as space runs out (or some other explicit rule applied such as via mmafmctl evict). Daniel Daniel Kidger HPC Storage Solutions Architect, EMEA [email protected]<mailto:[email protected]> +44 (0)7818 522266 hpe.com<http://www.hpe.com/> [cid:25007e89-6cb9-44a6-b342-34415f14c0ed] ________________________________ From: gpfsug-discuss <[email protected]> on behalf of Christian Petersson <[email protected]> Sent: 12 May 2022 13:47 To: IBM Spectrum Scale <[email protected]> Cc: gpfsug main discussion list <[email protected]> Subject: Re: [gpfsug-discuss] TCT Premigration Question Thanks for the answer, AFM to Cloud Object Storage did we look at but had some thoughts about. The base design was to replicate a copy from our primary cluster to the backup cluster where we push that data directly to S3. Our second cluster has much less disk space than our primary cluster and our primary cluster has a HSM function enabled to a tape library. If I start using AFM to Object, will that delete the file from my disk and only keep a stub file or similar on the disk? This is why we are looking at TCT instead of AFM Object, we don't have enough disk space to keep multiple copies of the data on disk, we need to use other storage layers like S3 and Tape. /C On Thu, 12 May 2022 at 14:38, IBM Spectrum Scale <[email protected]<mailto:[email protected]>> wrote: Some thoughts to consider regarding your question. 1 If this is “new” deployment, it is recommended that you evaluate the AFM object support since the TCT feature is stabilized and no future changes/enhancements will be made to it. 2 TCT doesn’t support Ceph cloud object storage. So, though it is an S3 compliant interface, and things are working, in case of any issues in future, IBM would require that you reproduce any problemin a supported environment. 3 If a file is skipped for any errors, if it is policy driven (pre)migration, the next policy run would pick up the file again. Even within a single (pre)migrate, TCT does 5 retries, on failure. So at both levels (within TCT as well as at policy level) file will be retried for migration. Regards, The Spectrum Scale (GPFS) team ------------------------------------------------------------------------------------------------------------------ If you feel that your question can benefit other users of Spectrum Scale (GPFS), then please post it to the public IBM developerWroks Forum at https://www.ibm.com/developerworks/community/forums/html/forum?id=11111111-0000-0000-0000-000000000479<https://www.ibm.com/developerworks/community/forums/html/forum?id=11111111-0000-0000-0000-000000000479>. If your query concerns a potential software error in Spectrum Scale (GPFS) and you have an IBM software maintenance contract please contact 1-800-237-5511 in the United States or your local IBM Service Center in other countries. The forum is informally monitored as time permits and should not be used for priority messages to the Spectrum Scale (GPFS) team. From: "Christian Petersson" <[email protected]<mailto:[email protected]>> To: [email protected]<mailto:[email protected]> Date: 05/11/2022 03:34 AM Subject: [EXTERNAL] [gpfsug-discuss] TCT Premigration Question Sent by: "gpfsug-discuss" <[email protected]<mailto:[email protected]>> ________________________________ Hi, I have set up a new Spectrum Scale Cluster where we are using TCT Premigration to a Ceph S3 Cluster, the premigration itself works fine and I got data to our S3 storage. But from time to time I reach our S3 max sessions and get the following Hi, I have set up a new Spectrum Scale Cluster where we are using TCT Premigration to a Ceph S3 Cluster, the premigration itself works fine and I got data to our S3 storage. But from time to time I reach our S3 max sessions and get the following errors in our policy. <1> MCSTG000130E: Command Failed with following reason: Request failed with error : com.ibm.gpfsconnector.messages.GpfsConnectorException: Unable to migrate data to the cloud, an unexpected exception occurred : com.amazonaws.ResetException: The request to the service failed with a retryable reason, but resetting the request input stream has failed. See exception.getExtraInfo or debug-level logging for the original failure that caused this retry.; If the request involves an input stream, the maximum stream buffer size can be configured via request.getRequestClientOptions().setReadLimit(int). But I wonder, if a file get skipt, will Spectrum Scale Automatic retry that file again later or will that be skipped and how can I catch does files for a new migration? -- Med Vänliga Hälsningar Christian Petersson E-Post: [email protected]<mailto:[email protected]> Mobil: 070-3251577 _______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at gpfsug.org<http://gpfsug.org> http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org<http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org> -- Med Vänliga Hälsningar Christian Petersson E-Post: [email protected]<mailto:[email protected]> Mobil: 070-3251577
_______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at gpfsug.org http://gpfsug.org/mailman/listinfo/gpfsug-discuss_gpfsug.org
