[petsc-users] petsc install

[Dharmendar Reddy]

On Thu, Feb 6, 2014 at 4:03 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
>
> On Feb 6, 2014, at 3:43 PM, Matthew Knepley <knepley at gmail.com> wrote:
>
>> On Thu, Feb 6, 2014 at 3:35 PM, Dharmendar Reddy <dharmareddy84 at gmail.com> wrote:
>> Hello,
>>          I use the petsc next via git to work on my local machine
>> (windows). Now i need to move my work to the server (linux) which has
>> no internet access ( i can talk to bitbucket from the server).
>
>    If the server has no internet access then how can you talk to bitbucket from the server?
>
>    Presumably the server has "no internet access" because of some misguided concern about security, yet ftp which is the least secure mechanism in history is used to copy files to this machine? And what is this ftp running on if it is not ftp?

Well, I can only ftp in only if i am on the vpn. But yes, its a
security concern.  I was thinking, i can maintian a local clone on my
windows machine and patch the server version when ever i do sync to
petsc-next on the local machine. I do not sync the code that often.

The plan may work only if i can download the external packages onto
local machine which is a windows system with no compilers, python
etc...

>
>    Barry
>
>> The
>> only way i can transfer files between the machines is via ftp from
>> local machine to linux server.
>>
>> 1. ) I want to download the external packages on the local machine and
>> transfer the files to the server for installation. Is there a
>> configuer flag which will just download the external packages used by
>> petsc ?
>>
>> You can always use --download-<package>=<path/to/tarball>. We do not have a flag
>> that just downloads tarballs.
>>
>> Satish, can you just use an SSH tunnel for this?
>>
>> 2.) How do i keep my server copy in sync with petsc-next ? when ever i
>> want to do a pull.
>>
>> I would just point it at a repo on the server, and keep that one up to date.
>>
>>    Matt
>>
>> Thanks
>> Reddy
>>
>>
>>
>> --
>> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.
>> -- Norbert Wiener
>