On 20 May 2015, at 17:01, Mindy <[email protected]> wrote: > > Good news or bad news, depending on how you slice it: running the extremely > naive test script below: > > ``` > for i in `seq 1 2000`; do curl -1 -k https://192.168.3.2 >/dev/null; done > ``` > > against a unikernel generated with mirage-seal results in the following:
That's good news from my front, since it's easily reproducible! Probably a good time to switch over to an issue on https://github.com/mirage/mirage/issues and take it from there. Dave: did you ever get a chance to look at the block device naming issue that was hindering adding support for dynamic storage for Xen? It's probably a good time to switch to dynamically loading in content for mirage-www and mirage-decks (with a buffer cache of course). That would bring the required memory back down for the default case of mirage-seal. I'll take a look if you haven't got time atm. -anil _______________________________________________ MirageOS-devel mailing list [email protected] http://lists.xenproject.org/cgi-bin/mailman/listinfo/mirageos-devel
