https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-13848 Lustre <p>Explanation from email:</p> <p>“ rwid=2 is the attach ID, each PCC backend has an attach ID to Uniquely identify itself on the client. For PCC-RW, it is usually same as the archive ID of the copy tool running on the client. The PCC-RO attach ID (roid) is same as PCC-RW attach id (archive ID) by default. “</p> <p>So there is 2 cases with HSM enabled and without HSM.</p> <ul> <li>Is it correct that  HSM acrhive ID  will be assigned by copytool?</li> </ul> <p>Second case need additional explanation for me.</p> <ul> <li>How to choose archive ID without HSM</li> <li>What is possible range for archive ID?</li> <li>Does user need to know anything else about archive ID? How it used internally?</li> <li>It is still unclear for me should be archive  ID the same on all client nodes or it should be different for PCC RO?</li> </ul> LU-13848

PCC RO documentation options needs additional explanation for user how to use them should added

Bug Trivial Open Unresolved Qian Yingjin Vikentsi Lapa Mon, 3 Aug 2020 09:30:46 +0000 Fri, 7 Aug 2020 06:50:19 +0000 0 3 <p>"Is it correct that HSM acrhive ID will be assigned by copytool?"<br/> Yes, the archive ID will be assigned and set when launch lhsm_posix copytool. It is used to uniquely identify the HSM backend.</p> <p>Second case need additional explanation for me.<br/> "How to choose archive ID without HSM"<br/> If a PCC backend is used for PCC-RW, then the RW attach ID (rwid) must be same as the archive ID of the copytool running on this client. As archive ID is used to uniquely identify the HSM backend in cluster wide, the archive IDs are usually not same. <br/> If a PCC backend only provides PCC-RO caching services without HSM, then the RO attach ID (roid) only needs to be unique on the client to identify the HSM backend. Different clients can use same RO attach ID.</p> <p>"What is possible range for archive ID?"<br/> archive ID in HSM is a 64 bit value, It is large enough, I think.</p> <p>"It is still unclear for me should be archive ID the same on all client nodes or it should be different for PCC RO?"</p> <p>For PCC-RO, the attach ID can be same on all client nodes if each client only has one PCC backend.<br/> But if a client has config two PCC-RO backend, then the attach ID for each PCC-RO backend on this client must be different and unique to identify PCC-RO backend on this client.</p> <p>"Is it correct that HSM acrhive ID will be assigned by copytool?"<br/> Yes, the archive ID will be assigned and set when launch lhsm_posix copytool. It is used to uniquely identify the HSM backend.</p> <p>Second case need additional explanation for me.<br/> "How to choose archive ID without HSM"<br/> If a PCC backend is used for PCC-RW, then the RW attach ID (rwid) must be same as the archive ID of the copytool running on this client. As archive ID is used to uniquely identify the HSM backend in cluster wide, the archive IDs are usually not same. <br/> If a PCC backend only provides PCC-RO caching services without HSM, then the RO attach ID (roid) only needs to be unique on the client to identify the HSM backend. Different clients can use same RO attach ID.</p> <p>"What is possible range for archive ID?"<br/> archive ID in HSM is a 64 bit value, It is large enough, I think.</p> <p>"It is still unclear for me should be archive ID the same on all client nodes or it should be different for PCC RO?"</p> <p>For PCC-RO, the attach ID can be same on all client nodes if each client only has one PCC backend.<br/> But if a client has config two PCC-RO backend, then the attach ID for each PCC-RO backend on this client must be different and unique to identify PCC-RO backend on this client.</p> Development Rank 1|i016u7: Rank (Obsolete) 9223372036854775807 Severity