summaryrefslogtreecommitdiff
path: root/open_issues/cloud.mdwn
diff options
context:
space:
mode:
Diffstat (limited to 'open_issues/cloud.mdwn')
-rw-r--r--open_issues/cloud.mdwn49
1 files changed, 49 insertions, 0 deletions
diff --git a/open_issues/cloud.mdwn b/open_issues/cloud.mdwn
new file mode 100644
index 00000000..58ed2f5b
--- /dev/null
+++ b/open_issues/cloud.mdwn
@@ -0,0 +1,49 @@
+[[!meta copyright="Copyright © 2013 Free Software Foundation, Inc."]]
+
+[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
+id="license" text="Permission is granted to copy, distribute and/or modify this
+document under the terms of the GNU Free Documentation License, Version 1.2 or
+any later version published by the Free Software Foundation; with no Invariant
+Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license
+is included in the section entitled [[GNU Free Documentation
+License|/fdl]]."]]"""]]
+
+Some *cloud*y things.
+
+[[!toc]]
+
+
+# [[!wikipedia OpenStack]]
+
+## IRC, freenode, #hurd, 2013-09-21
+
+ <jproulx> Hmmm, was hoping to run hurd on my kvm based openstack cloud, but
+ no virtio.
+ <jproulx> I see "Write virtio drivers for KVM. Ideally they would be
+ userland" is listed as a "small hack", as a sysadmin rather than an OS
+ hacker it doesn't sound small to me, but if there's some standard
+ documentation on porting drivers I could take a run at it.
+ <youpi> well, perhaps "small" is not the proper word
+ <youpi> compared to e.g. revamping disk i/o :)
+ <youpi> it's not something one can achieve in e.g. 1h, for instance
+ <youpi> it's not something straightforward either, one has to get
+ documentation about virtio (I don't know what exists), and get
+ documentation about the mach device interface (that's in the gnumach
+ manual, the devnode translator can be used as a skeleton)
+ <youpi> jproulx: openstack imposes the use of virtio drivers? that's odd
+ <jproulx> that's more like I'd expect. I there's enough search terms in
+ your response for me to see what's really involved
+ <jproulx> youpi it doesn't impose that but it is how mine is configured the
+ other thousand VMs are happier that way.
+ <jproulx> I can look at that side too and see if I need to have everything
+ use the same device settings or if I can control it per instance
+ <jproulx> A bit of a non-sequitur at this point but just in case someone
+ searches the transcripts and sees my questions about hurd on openstack,
+ yes it is possible to specify non-virtio devices per image, here's the
+ commandline to load sthibault's qemu image into openstack with devices
+ that work:
+ <jproulx> glance image-create --property hw_disk_bus=ide --property
+ hw_cdrom_bus=ide --property hw_vif_model=rtl8139 --disk-format raw
+ --container-format bare --name gnu-hurd --copy-from
+ http://people.debian.org/~sthibault/hurd-i386/debian-hurd.img
+ <youpi> jproulx: thanks, I've pushed it on the wiki