Opened 16 years ago
Closed 10 years ago
#4064 closed defect (obsolete)
Internal network very slow gPXE booting a VM from another (Linux) VM
| Reported by: | James Mansion | Owned by: | |
|---|---|---|---|
| Component: | network | Version: | VirtualBox 2.2.2 |
| Keywords: | PXE, gPXE | Cc: | |
| Guest type: | Linux | Host type: | Windows |
Description (last modified by )
I am trying to develop a system where I load gPXE's undionly.kpxe via TFTP and then complete the process using HTTP under control of gPXE.
Normally it gets gPXE very quickly, gets the tiny script, and then gets the next images very slowly - about 14 or so 1518 byte packets per second. This rate seems independant of the NIC used in the client VM and also independant of the web server in use. The boot server is Xubuntu 9.04.
Sometimes (very infrequently) the performance is much, much faster.
Debugging flags in gPXE do not show any errors being detected and the packet size is consistent.
Attachments (1)
Change History (4)
by , 16 years ago
comment:1 by , 16 years ago
I tried to upload a pcap for booting to load just undionly.kpxe, a tiny script, and then (with http) a Linux kernel (no initrd) - but its 6MB and even after bzip2 its still far too big for the maximum upload.
comment:2 by , 16 years ago
Huh, I can't stype, its http://barnhouse.me.uk/file.pcap.bz2
Sorry about that.
comment:3 by , 10 years ago
| Description: | modified (diff) |
|---|---|
| Resolution: | → obsolete |
| Status: | new → closed |
Log file for the boot client