svn commit: r1396893 - /incubator/tashi/board/2012-10.txt

[st...@apache.org]

Author: stroucki
Date: Thu Oct 11 02:11:29 2012
New Revision: 1396893

URL: http://svn.apache.org/viewvc?rev=1396893&view=rev
Log:
Convert commits to a few condensed paragraphs.


Modified:
    incubator/tashi/board/2012-10.txt

Modified: incubator/tashi/board/2012-10.txt
URL: http://svn.apache.org/viewvc/incubator/tashi/board/2012-10.txt?rev=1396893&r1=1396892&r2=1396893&view=diff
==============================================================================
--- incubator/tashi/board/2012-10.txt (original)
+++ incubator/tashi/board/2012-10.txt Thu Oct 11 02:11:29 2012
@@ -15,49 +15,26 @@ physical aspects of a cluster like power
 handing out physical machines.
 
 Activities July-October:
-tashi-client: leave ValueErrors which came from incorrect operations, while converting ones originating from a failure to perform a proper operation to TashiError
-tashi-client: Don't tell the user about an "exception"; offer help if a ValueError occurred.
-improve error reporting to the user
-tashi-client: check for permissions when doing destroyMany or shutdownMany
-qemu: Address race condition on instantiation, where the cleanup
-thread would remove a VM being born. Allow 5 seconds for the launch of
-the kvm process.
-improve messages given to the user
-Do the trivial addition of setHostNotes for now. "Real Soon Now", this type of function will be put into an administration tool.
-import Diego Gomes' improvements to detecting guest IP addresses, and adjust documentation. Thanks Diego!
-replace deprecated with current functions
-now have a dedicated cluster administration client command, supporting host state and comment changes, user management and network management. There were removed from the regular tashi client
-qemu.py: Raise EOFError rather than RuntimeError when reading off the monitor FD
-qemu.py: Handle EOFError in processVmStats to avoid a stack dump when a VM is exiting
-qemu: replace the fork and shell and head call to get system memory with a python only solution
-qemu: make logging the same as in other modules
-allow automatic host registration with configuration option
-tashi-client: reraise network exceptions so we can handle them
-database: make sure for new hosts, host structure is fully initialized
-qemu: make note on hostIds, and the need make sure they match ours in case we accept a migrated instance
-qemu: set state to running to avoid persistent states of MigrateTrans from the origin host when the NM restarts
-qemu: we ask for our hostname a lot, lets just do it once
-qemu: force started instances to have my hostId
-nodemanagerservice: correct hostId's when we get info for the CM
-tashi-admin: allow hosts to be reserved for a user
-
 
 In the period from July to October, the project did not ask to make 
 another incubating release, but is ready to start the process for a new 
 release incorporating the development efforts of the last 6 months.
 
-Development efforts this period have included providing a separate
-administration client, ...
+Development efforts this period have included providing a separate 
+administration client, allowing addition of users and networks, and host 
+reservations and availability for scheduling.
 
 The project has received code contributions from one non-committer in 
-this period. Diego Gomes...
+this period. Diogo Gomes provided support for deriving the IP addresses 
+of guests automatically, without having to scan the subnet. Thanks Diogo!
 
+Additional stability and user experience improvements were also 
+committed.
 
-Upcoming software goals are 
-to investigate what is needed to 
-support IPv6, replace RPyC, and to provide the ability to hand out 
-server slices (operating system level virtualization). Besides CPU and 
-memory, disk storage should also be a schedulable resource.
+Upcoming software goals are to investigate what is needed to support 
+IPv6, replace RPyC, and to provide the ability to hand out server slices 
+(operating system level virtualization). Besides CPU and memory, disk 
+storage should also be a schedulable resource.
 
 The project has a user community, but it is small. Growth mostly has 
 happened by word of mouth. To show potential users at large the utility