1 |
I am in exactly the same shoes as you. |
2 |
|
3 |
On Feb 15, 2006, at 4:59 AM, Timo Veith wrote: |
4 |
|
5 |
> Am Mittwoch 15 Februar 2006 01:02 schrieb Michael Stewart (vericgar): |
6 |
>> You really should set up a test environment (if you don't have one |
7 |
>> already) and do it there. |
8 |
> |
9 |
> Hmm, I've done a upgrade on my personal workstation and it went ok. |
10 |
> So the |
11 |
> issue is not that I haven't done any gcc upgrade, yet. But on my |
12 |
> workstation nobody else is affected if anything goes wrong. I can do |
13 |
> reboots, unmerges or remerges etc. |
14 |
> |
15 |
> I am a little concerned about doing it on the production server, |
16 |
> although |
17 |
> I don't know of any problems yet. I just want to be carefull. |
18 |
> |
19 |
>> You may also want to consider setting up a build server which will |
20 |
>> compile your packages for your specific setup so all you have to |
21 |
>> do on |
22 |
>> the production server is merge the binary. |
23 |
> |
24 |
> Of course in general this is a good thing, but setting up a build |
25 |
> environment for just about five production systems seems to be too |
26 |
> much |
27 |
> overhead to me. The machines we have aren't the same, as far as |
28 |
> architecture is concerned. So I would have to build packages with |
29 |
> different compiler flags. Probably there is a mechanism to solve that |
30 |
> issue, but is it really worth the effort? I would hold the compilation |
31 |
> away from the production system, but I could do it on day times where |
32 |
> only few ressources are needed. Also I can set PORTAGE_NICENESS. |
33 |
> |
34 |
I have 6 production servers most with different archs. I feel there |
35 |
is a lot of overhead also. not to mention we are a start up and all |
36 |
of my servers are being used to the fullest. We at this point in time |
37 |
cant afford to have a test server. |
38 |
|
39 |
>> As far as the gcc upgrade, I haven't done it myself. Is there a |
40 |
>> specific reason you need the latest version? If not, then why |
41 |
>> upgrade? |
42 |
|
43 |
I like this approach however, my question is when will we be forced |
44 |
to upgrade. |
45 |
|
46 |
> |
47 |
> I have learned that it is not possible to stay at a certain level of |
48 |
> package versions. You need to follow the flow, if you want to install |
49 |
> security updates. I know there are projects that want to take care of |
50 |
> this issue, but I think they aren't ready, yet. So I believe it could |
51 |
> only be a matter of time until you have to upgrade gcc. |
52 |
> |
53 |
> Of course I could wait until it comes to that point in time. I am just |
54 |
> curious about what you think about and how you handle this. |
55 |
> |
56 |
|
57 |
When this point in time comes.. It is time to shed a tear for me.. I |
58 |
hope by then I get new servers and rotate services with fresh installs. |
59 |
|
60 |
> -- |
61 |
> Timo |
62 |
> -- |
63 |
> gentoo-server@g.o mailing list |
64 |
> |
65 |
|
66 |
-- |
67 |
gentoo-server@g.o mailing list |