On Wed, Nov 18, 2015 at 1:05 PM, Per Jessen <per@computer.org> wrote:
I'm surprised Ironport can work that out, it's fairly complex. You're certain it does that? Ah, maybe it does indeed fetch the whole file first, but that would be ridiculous. If a 16Mb file is split into 16 segments, that could potentially cause 16 full downloads :-)
Yep. The reason it does this is that it wants to scan the item for virus. I guess it needs the whole file for this. I recall something about information in the HTTP commands generated by zypper that indicate this is what is happening. I think that zypper was doing the right thing. It was the Cisco device that did not get this correct. Either in it's replies or in intercepting the HTTP commands used by libcurl. If memory serves...
Aha, that's interesting. zypper is certainly not the only one to use the segmented download, but the only other one I can think of right now is youtube.
Yep. Other clients had problems, which was how it was discovered to be an Ironport issue.
Anyway, yes, it sounds like the ZYPP_MULTICURL option is in fact the solution you need.
Except it seems not to be working well with big files, as it had been doing previously. Of course, maybe the IT guys here have changed something as well. They can be rather secretive about these things. Moving targets everywhere... -- Roger Oberholtzer -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org