what has 'yum update' done?

Reindl Harald h.reindl at thelounge.net
Sun Jul 14 13:22:55 UTC 2013


Am 14.07.2013 15:12, schrieb Frank Murphy:
> On Sun, 14 Jul 2013 14:55:02 +0200
> Reindl Harald <h.reindl at thelounge.net> wrote:
> 
> Didn't know wherer you wanted to post back to list?
> If yes, I will.

i had the list also as RCPT a well as now
so no need to re-post

it makes pretty no sense having a discussion where
answers are releaed Tuesday because moderation and
until then are out of context

>> Am 14.07.2013 14:41, schrieb Frank Murphy:
>>> On Sun, 14 Jul 2013 00:51:05 +0200
>>> lee <lee at yun.yagibdah.de> wrote:
>>>
>>>> Why not make it one of the great features of Fedora?  Then take
>>>> it to the next step and make systemd send mails in case a service
>>>> couldn't be started or has issues,
>>>
>>> You can do that yourself currently
>>
>> and do this automatically is a bad idea
> 
> But, "lee" asked for it as feature, no need,
> is what was I was answering.

and i added information

>> how often should the ybe sent if a service goes crazy
>> and starts crashing 5 seconds after start, is restarted
>> by "Restart=always" and flooding your inbox - such behavior
>> is known as "self DOS" if implemented not 100% perfect
> 
> Not, if "lee" don't use Restart=Always,
> and would like to know if service x fails,
> so he can then have a look at it.

you do not know which servic es are using it in the future
after the damage is done it's to let to think about the impact

>> i have scripts parsing /var/log/messages and sent out notifies
>> every 30 minutes if something went wrong, all important services
>> have Restart=always and so i only need to know that it
>> happened and how often it happens
> 
> Knowing it broke, is not always enough,
> "lee" might want to  debug it.
> Maybe he's a developer?, 
> he hasn't provided that info as yet

and after i have my notify that "service X" crashed 5 times in
the last 30 minutes i can also start debugging, if i see
soemthing crashs too often and if i have yet time to debug

that's why it should be good considered *when* and *how often*
you get so notifies, it's not easier to debug if you are at it
and you get each minute a new notify nor does it help much
having in the morning 100 messages saying "service X has a problem"

that's all about rate-limiting and collectiong relevant infos
without too much noise all day long - if it get's too much
you simply start ignore the notifies leading to ignore
sooner or later a really imporant one


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 263 bytes
Desc: OpenPGP digital signature
URL: <http://lists.fedoraproject.org/pipermail/users/attachments/20130714/a1564465/attachment.sig>


More information about the users mailing list