Hi Peter,
> BDII, CREAM, DPM mysql, DPM disk (3 of them), TORQUE_server
> How long do you think it will take to provision and setup the OS layer and then the middleware layer for all?
There is no way to tell - it depends on many factors. I can't say much about DPM mysql and DPM disk. But I can say things about BDII, CREAM, and TORQUE_server.
Each one builds on the other. Once the hardware/OS/network layers are sorted for (say) BDII, then those layers are also sorted for
CREAM and TORQUE_server (and (to a far less extent) DPM mysql and DPM disk). They are similar. We use kickstart and puppet for these things, and it would take N days to port it from that to some other, manual "build system". Then you'd have to test it and compare it (how long is a piece of string)?
But, once you have built and commissioned (say) the BDII (which takes few resources and can be easily hosted on some cheap machine?), you would be able to tell how long it will take to build the hardware/OS/network layers for the other systems. And it will also give you a clue about the process for the other node types. So, you could use the BDII to (partly) calibrate the process for the other systems. You can have two site BDII's at once, no problem, and you can compare their contents to make sure they are functionally identical.
With respect to build times for CREAM, and TORQUE_server; these are hard to estimate. It was necessary here at Liverpool to build a CREAM, and TORQUE_server and an lcg-CE in a period of downtime a year or so ago. It was a week of early starts and late nights before all that was working properly together. But I wasn't particularly clued up about it. The "Big Bang" approach (that you are talking about) was fearsome to me, and I'd always try to take an incremental approach if I could. The Big Bang approach is basically the same as a total disaster recovery, but without the experience of having already gone through the build process ever before!
With respect to the primary source of install info (GenericInstallGuide320). It's a good place to start, but I personally had to be quite "creative" compared to the standard process described therein. The guide will take you part of the way there, but there's always plenty more to do, and some of the information is poorly presented.
In short - incremental is the best way, but if you must go "Big Bang", then have lots of tea/coffee on tap - you'll need it.
Steve
Peter Grandi wrote:
> About switching to SL5, essentially from scratch, I am about to install alongside our SL4/gLite 3.1 server nodes a set of equivalent SL5/gLite 3.2 ones, with node types:
>
>
> The APEL and UI servers and the WNs are already on SL5.
>
> I may be able to copy several configuration files across from the SL4 version to the SL5 one, but the SL4/gLite 3.1 systems have a number of hand-made somewhat hacked configurations.
>
> Each service will be installed in its own SL5 VM, except for the DPM disk servers which will be on real hw.
>
> The installation will be manual (no QUATTOR/Cfengine/..., just install and then sort out later).
>
> Assuming I do nothing else and work nominal hours (neither will happen :->, but I need a baseline):
>
> * How long do you think it will take to provision and setup the OS layer and then the middleware layer for all?
>
> * Additionally to commission the lot into a smoothly running system?
>
> Because there is the possibility of just declaring a site downtime and then at risk time for the duration, to avoid having to maintain the old set, and to avoid having both sets of VMs running at the same time on the same hosts.
>
> Any suggestions, shortcuts, tips welcome. My main sources will be the main installation page:
>
> > https://twiki.cern.ch/twiki/bin/view/LCG/GenericInstallGuide320
>
> and the (always useful) ScotGrid (and others) blogs.
>
> I am tempted to be brave and install EMI 1 instead of gLite 3.2, as the latter will be obsoleted in around 10 months time, but I wonder how risky it would be.
>
--
Steve Jones [log in to unmask]
System Administrator office: 220
High Energy Physics Division tel (int): 42334
Oliver Lodge Laboratory tel (ext): +44 (0)151 794 2334
University of Liverpool http://www.liv.ac.uk/physics/hep/
|