Sharing cache in a computing system
View Patent ↗Sharing cache in a computing system that includes a plurality of enclosure attached servers, including: identifying, by an enclosure, a first enclosure attached server that is not meeting a first predetermined performance threshold; identifying, by the enclosure, a second enclosure attached server that is meeting a second predetermined performance threshold; blocking, by the enclosure, access to a predetermined amount of cache on the second enclosure attached server by the second enclosure attached server; determining, by the enclosure, whether the second enclosure attached server is meeting the second predetermined performance threshold; responsive to determining that the second enclosure attached server is meeting the second predetermined performance threshold, lending, by the enclosure, the predetermined amount of cache on the second enclosure attached server to the first enclosure attached server.
1. A method comprising:
identifying a first server that is not meeting a first predetermined performance threshold;
identifying a second server that is meeting a second predetermined performance threshold, wherein identifying the second server that is meeting the second predetermined performance threshold comprises:
querying a plurality of servers for performance information,
receiving performance information from each of the plurality of servers, and
comparing the performance information from each of the plurality of servers to predetermined performance thresholds for each of the plurality of servers;
blocking access to a predetermined amount of cache on the second server by the second server;
determining after blocking access by the second server to the predetermined amount of cache on the second server,
whether the second server is meeting the second predetermined performance threshold;
responsive to determining that the second server is meeting the second predetermined performance threshold, lending the predetermined amount of cache on the second server to the first server.
2. The method of claim 1 further comprising, responsive to determining that the second server is not meeting the second predetermined performance threshold, restoring access to the predetermined amount of cache on the second server by the second server.
3. The method of claim 1 wherein blocking access to the predetermined amount of cache on the second server by the second server further comprises identifying the least frequently used portions of the cache on the second server.
4. The method of claim 1 further comprising:
identifying that the predetermined amount of cache on the second server should be returned to the second server;
unlending the predetermined amount of cache on the second server to the first server; and
restoring access to the predetermined amount of cache on the second server by the second server.
5. The method of claim 4 wherein identifying that the predetermined amount of cache on the second server should be returned to the second server further comprises identifying that the first server does not need the predetermined amount of cache on the second server.
6. The method of claim 4 wherein identifying that the predetermined amount of cache on the second server should be returned to the second server further comprises identifying that the second server needs the predetermined amount of cache on the second server.
7. The method of claim 1 wherein identifying, by the enclosure, the second server that is meeting a second predetermined performance threshold further comprises polling each of the plurality of servers for performance information.
8. The method of claim 1 , wherein the first server and the second server do not use the predetermined amount of cache on the second server during the blocking.
9. The method of claim 1 further comprising accessing, by the first server, the predetermined amount of cache on the second server indirectly via an enclosure.