Eli Stevens's questions

I am trying to find a metric to use that will tell me if starting another process estimated to use N bytes of memory will have an adverse impact on system performance.

I know that the /proc/meminfo fields MemFree, Buffers, and Cached can be added together to give the total memory available to programs. However, I would prefer to not squeeze out cache that's actually being used to good effect (the workload results in disk hotspots).

The Active entry in /proc/meminfo seems like it could be compared with MemTotal to determine how much memory is available (either by swapping something idle out, or by being free in the first place), but I have not been able to find much information about Active and so I'm not sure.

Should I aim to have Active memory as close to MemTotal without going over as I can manage?

Is there some other way to estimate the memory behavior of running another process?

TLDR: for insurance purposes, we want to alert the recipient of a shipped assembled computer that there might be damage to the components via a shockwatch label. What G-force rating should we use?

We are having problems moving assembled computers via standard shipping services due to rough handling causing damage to the internal components (specifically the GPUs are snapping the PCI-E slots off of the motherboard). Having the shipper provide the outer layer of packaging hasn't helped.

Insurance requires that the unboxing process be documented, but it can be difficult to tell that documentation will be required without opening the packaging, so we want to apply shockwatch labels to the external packaging to alert the recipients.

However, it's not clear what rating we should use. We don't have the equipment to determine this ourselves (nor do we have the desire to destroy thousands of dollars worth of hardware to determine it experimentally). The carriers, obviously, don't make any claims about things like maximum drop height, etc. and so far, the only shock rating information we've found online has been specifically for drives (which haven't experienced damage so far).

We don't need to be very precise about the shock rating, just low enough to catch most of the drops that will cause damage, and high enough to not trigger for every shipped machine. We are also hoping that the presence of a shockwatch label will encourage less cavalier treatment of the package.

Obviously, we are also iterating on our packaging to reduce the likelihood of damage.

After explaining our situation to a sales rep from Shockwatch, he said that typical handling of a 50 lb. package will cause the 25G to trigger every time, and suggested the 50G label.

I'm going to have an Ubuntu 11.10 box in a customer's data center running a custom webapp. The customer will not have ssh access to the box, but will need authentication and authorization to access the webapp. The customer needs to have the option of either pointing the webapp at something that we've installed locally on the machine, or to use an Active Directory server that they have. I plan on using a standard "users belong to groups; groups have sets of permissions; the webapp requires certain permissions to respond" auth setup.

What software stack can I install locally that will allow an easy switch to and from an Active Directory server, while keeping the configuration as simple as possible (both for me and the end customer)?

I would like to use as much off-the-shelf software for this as possible; I do not want to be in the business of keeping user passwords secure. I could see handling the user/group/permission relationships myself if there is not a good out-of-the-box solution (but that seems highly unlikely).

I will accept answers in the form of links to "here is what you need" pages, but not "here is what Kerberos does" unless that page also tells me if it's required for my use case (essentially, I know that AD can speak Kerberos, but I can't tell if I need it to, or if I can just use LDAP, or...).