Camelia, the Perl 6 bug

IRC log for #etools, 2012-07-22

| Channels | #etools index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:07 mikemol joined #etools
01:07 mikemol oi
01:08 mikemol Had hardware failures compound for so long, it took reading my own README file before I remembered this IRC channel.
01:08 mikemol Just pushed a bunch of cleanups to git. Trying to get things cleaned up enough that I can do a release, and then go on to work on entmesh/
01:19 snadge i've been interested in seeing that
01:19 snadge at work we use quantis pci hardware rng
01:20 snadge and xen virtualised hosts.. and there is no clean way of getting entropy from the dom0 (host) into the domUs (vm guests)
01:21 snadge there is a program called entropy broker.. but it hasn't been updated in years, and its kinda dodgy
01:22 snadge the other solutions i was looking at, involve basically cating the random device over a tcp socket.. and then piping that into rngd.. which would work, but seems also a bit flakey
01:26 mikemol snadge: You're welcome to lend a hand with the entmesh authoring. I'd like to tackle it by splitting entbuf and entwash up such that they mostly use common code.
01:26 mikemol s/entwash/entwatch/
01:28 * mikemol writes an ebuild for etools.
04:46 snadge yeah im not overly interested in being involved with development of a solution, being a sysadmin, and outside of the scope of what im required/can be bothered to do ;)
04:47 snadge the other sysadmin suggests migrating to kvm, which has virtio-rng
04:47 snadge but my apathy / laziness is also against that solution.. because its something i need to look into and understand anyway
04:48 snadge ideally what im after is something i can just put in place quickly, be fairly assured its not going to fail.. and forget about
04:51 snadge and have some degree of confidence that its doing the job that its intended to do
05:42 richardus joined #etools
05:42 mikemol joined #etools
05:42 snadge joined #etools
12:32 mikemol snadge: entmesh is going to be pretty experimental for a while, even after all of the obvious code is written. And given the peer-to-peer nature of it, I fully expect some strange behaviors to present themselves if people aren't attentive when they set things up.
12:34 mikemol As an example, if you don't explicitly set up some kind of one-way flow, it's very probable that, under some circumstances, you'll get an oscillating pattern where entropy is pushed back and forth between small groups of nodes which are alternately flush with and starved of entropy.
22:59 nb joined #etools

| Channels | #etools index | Today | | Search | Google Search | Plain-Text | summary