The only v2 difference is - do not call sethostname if new name is empty.
- Original Message -
> From: "Justin Cinkelj"
> To: osv-dev@googlegroups.com
> Cc: "Justin Cinkelj"
> Sent: Thursday, November 17, 2016 4:06:53 PM
> Subject: [PATCH v2] cloud-init: use hostname from yaml file
>
>
Previously cloud-init set VM hostname if EC2/GCE was used as source. But
it ignored hostname if it was specified via yaml file or
contextualization image. The commit fixes this, so that hostname can be
set via yaml file too.
The example cloud-init.yaml file is updated with commented-out examples
Previously cloud-init set VM hostname if EC2/GCE was used as source. But
it ignored hostname if it was specified via yaml file or
contextualization image. The commit fixes this, so that hostname can be
set via yaml file too.
The example cloud-init.yaml file is updated with commented-out examples
Thank you, for accepting and improving.
Justin
- Original Message -
> From: "Nadav Har'El"
> To: "Justin Cinkelj"
> Cc: "Osv Dev"
> Sent: Thursday, November 17, 2016 1:03:56 PM
> Subject: Re: [PATCH] dhcp: propose our hostname to DHCP server
>
> Thanks. I pushed this patch, but added a
在 2016年11月17日星期四 UTC+8下午7:13:38，Joe Duarte写道：
> On Wednesday, November 16, 2016 at 5:02:32 AM UTC-7, Pekka Enberg wrote:On
> Wed, Nov 16, 2016 at 9:18 AM, wrote:
>
> > Unikernel has a lot of advantages(fast,small,secure),it is a good
> > choice to be used on mobile applications.Recently,
Thanks. I pushed this patch, but added another patch: Do not call
sethostname() if the DHCP server did not set the host name. Let OSv
continue to use the silly default it has ("osv.local").
--
Nadav Har'El
n...@scylladb.com
On Wed, Nov 16, 2016 at 8:42 PM, Justin Cinkelj
wrote:
> Send our host
From: Nadav Har'El
Committer: Nadav Har'El
Branch: master
dhcp: ignore empty hostname
If the DHCP server does not send back a hostname, the existing code
called sethostname() with an empty host name. Do not do that.
Signed-off-by: Nadav Har'El
---
diff --git a/core/dhcp.cc b/core/dhcp.cc
--
From: Justin Cinkelj
Committer: Nadav Har'El
Branch: master
dhcp: propose our hostname to DHCP server
Send our hostname to DHCP server. This gives DHCP server option to
update DNS server.
The usecase assumes that hostname will be set to descriptive/unique
value (say via cloudinit) before DHCP
On Wednesday, November 16, 2016 at 5:02:32 AM UTC-7, Pekka Enberg wrote:
>
> On Wed, Nov 16, 2016 at 9:18 AM, >
> wrote:
> > Unikernel has a lot of advantages(fast,small,secure),it is a good
> choice to be used on mobile applications.Recently,I was thinking about
> running Android appli
From: Nadav Har'El
Committer: Nadav Har'El
Branch: master
rng: ignore zero bytes sent by an entropy source
If we get zero bytes from an entropy source, we should obviously do
nothing with this input. But it's not clear that entropy_processor()
really ignores zero bytes as expected, so let's ex
From: Nadav Har'El
Committer: Nadav Har'El
Branch: master
virtio-rng: don't hang while waiting for entropy
random_kthread() has a loop, where it tries to get some entropy from
all sources (including the ring of internal events), feeds it into the
entropy pool, and then sleeps a bit and tries a
On Thu, Nov 17, 2016 at 12:23 PM, Gregor Berginc
wrote:
> Nadav, it does solve the problem with the delay. The service starts
> immediately after the instance is launched when using virtio-rng.
>
>
Excellent. It was your observation that nailed this bug: We did have enough
internal events to set
Nadav, it does solve the problem with the delay. The service starts immediately
after the instance is launched when using virtio-rng.
> On 17 Nov 2016, at 09:41, Nadav Har'El wrote:
>
> Gregor, can you please test if this patch really makes your /dev/urandom hang
> go away?
>
> Tomek and/or
On Thu, Nov 17, 2016 at 12:05 PM, Nadav Har'El wrote:
> If we get zero bytes from an entropy source, we should obviously do
> nothing with this input. But it's not clear that entropy_processor()
> really ignores zero bytes as expected, so let's explicitly ignore
> such input.
>
> Signed-off-by: Na
If we get zero bytes from an entropy source, we should obviously do
nothing with this input. But it's not clear that entropy_processor()
really ignores zero bytes as expected, so let's explicitly ignore
such input.
Signed-off-by: Nadav Har'El
---
bsd/sys/dev/random/live_entropy_sources.cc | 5 ++
On Thu, Nov 17, 2016 at 11:16 AM, Nadav Har'El wrote:
>
> On Thu, Nov 17, 2016 at 10:57 AM, Pekka Enberg wrote:
>>
>>
>> The virtio-rng part looks correct:
>>
>> Reviewed-by: Pekka Enberg
>>
>> However, perhaps we ought to also fix live_entropy_sources_feed()
>> _not_ to feed events to entropy_p
On Thu, Nov 17, 2016 at 10:57 AM, Pekka Enberg wrote:
>
> The virtio-rng part looks correct:
>
> Reviewed-by: Pekka Enberg
>
> However, perhaps we ought to also fix live_entropy_sources_feed()
> _not_ to feed events to entropy_processor() if entropy source read
> returned zero like later version
On Thu, Nov 17, 2016 at 10:39 AM, Nadav Har'El wrote:
> random_kthread() has a loop, where it tries to get some entropy from
> all sources (including the ring of internal events), feeds it into the
> entropy pool, and then sleeps a bit and tries again, ad infinitum.
>
> If the read function of *on
Usecase of this patch is for example HDFS. There we have one java
program, which can be run as (one) namenode or as (multiple) datanodes.
We whish to provide convinient shortcut for selection of the two distinc
java commandline parameters. So we put all required parameters into
two script-like file
Compared to v1 patch I only added additional patch to allow emtpy
command line. At least one application still needs to be specified,
but now it can be either via command line or via runscript option.
Justin Cinkelj (2):
loader: run commands from preconfigured scripts
loader: shutdown only if
Loader requires at least one program to be run via command line. With
runscript option, command line could be empty, and application to be run
would be specified via runscript only. The check for missing application
is modified to allow this case.
Signed-off-by: Justin Cinkelj
---
loader.cc | 11
Gregor, can you please test if this patch really makes your /dev/urandom
hang go away?
Tomek and/or Pekka, can you please review this patch? (its comment is much
longer than the actual patch ;-) ).
Thanks.
--
Nadav Har'El
n...@scylladb.com
On Thu, Nov 17, 2016 at 10:39 AM, Nadav Har'El wrote:
random_kthread() has a loop, where it tries to get some entropy from
all sources (including the ring of internal events), feeds it into the
entropy pool, and then sleeps a bit and tries again, ad infinitum.
If the read function of *one* of the entropy sources blocks, the entire
random_kthread() bl
23 matches
Mail list logo
