Saving result when running via pxeboot?

[aggieNick02]

We're interested in the site license's PXE booting capability, but we use iPXE so we have no way to upload the result files. We need to somehow get the results in order to automate things sufficiently. Can we specify a location to save the result files (e.g., a local disk or networks share ) or provide a web server url in the mt86.cfg file to upload the result file via a post? If not, is there a chance this might be added in the future, or can you recommend another approach to automate submission of results when using iPXE?

[David (PassMark)]

At the moment you can't write results to a local hard drive if doing a PXE network boot. (generally you do a PXE network boot because the local disk isn't formatted, or doesn't exist at all). Writing the report to network drive isn't possible, as there are no network device drivers loaded up in UEFI if you are doing a PXE boot.

Regarding the particular iPXE implementation of PXE: Their implementation of TFTP read/write was rather poor. There was no way to get the file size in order to do a TFTP read and TFTP write does not even exist. We believe this is due to iPXE first booting iPXE EFI binary, which chainloads MemTest86. If you could boot MemTest86 directly, we think the default EFI firmware can handle TFTP uploads. As this is all pretty complex, we would suggest using Serva PXE instead. Then you can get the test report written back to the PXE server.

ALSO, we can offer integration of MemTest86 results into a Management Console Database. If that is of interest, Email us for details.

[aggieNick02]

Thanks for the details David. Unfortunately moving to Serva PXE isn't an easy option for us, as we're running ipxe and dhcpd on linux right now. I see that Serva offers a proxyDHCP server so I can look into it a bit more, but we'd have to at the least add a windows box to our network and figure out if ServaPXE allows the per-machine customization that we currently have set up with our iPXE scripts.

If writing to a local drive ever becomes possible, that would be really useful to us. We would harvest the result file from the drive when the machine restarts post-MemTest86. We have enough control pre- and post- MemTest86 that we can format/partition/identify the drive (as there may be more than one local disk) in whatever way MemTest86 required. Thanks again.