[ipxe-devel] Could iPXE boot faster than local HDD.

[Andrew Bobulsky]

Hello Jerry,

I can't speak to the anomaly you're seeing with iscsi, but you /may/ be able
to get a significant improvement on boot times by using iPXE's AoE initiator
instead, as AoE is a much lighter weight protocol.

The Etherboot wiki has some fairly comprehensive AoE setup guides if you'd
like to try it out :)

Cheers,
Andrew Bobulsky

On May 22, 2011, at 9:53 PM, "jerrycheng-hinet" <jaspers.cheng at msa.hinet.net>
wrote:

Hi, all,

I tried to compare windows xp booting time by using Microsoft bootvis.

For the first glance, ipxe always slower than local HDD than about 5
seconds. The bootvis showed the disk init time of local HDD is about 2.2
seconds, and iPXE is about 15.5 seconds.

Than I tried to use wireshark to observe what happened in iPXE's disk init
time. From the following captured message, I found there has a 15 seconds
gap that iSCSI connection seemed to get into trouble.

29408 15.646359 192.168.1.63 192.168.1.2 iSCSI SCSI: Data Out LUN: 0x00
(Write(10) Request Data)
29409 15.646374 192.168.1.2 192.168.1.63 TCP iscsi-target > 28906 [ACK]
Seq=18548589 Ack=86633 Win=18760 Len=0 TSV=624524 TSER=577920
29410 15.646818 192.168.1.2 192.168.1.63 iSCSI SCSI: Response LUN: 0x00
(Write(10)) (Good)
29411 15.647008 192.168.1.63 192.168.1.2 TCP 28906 > iscsi-target [ACK]
Seq=86633 Ack=18548637 Win=8192 Len=0 TSV=577920 TSER=624524
29412 17.443849 QuantaCo_9b:e4:69 Broadcast ARP Gratuitous ARP for
192.168.1.63 (Request)
29413 18.018077 QuantaCo_9b:e4:69 Broadcast ARP Who has 192.168.1.2?  Tell
192.168.1.63
29414 18.018090 Giga-Byt_d6:86:bc QuantaCo_9b:e4:69 ARP 192.168.1.2 is at
1c:6f:65:d6:86:bc
29415 18.018191 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [SYN] Seq=0
Win=65535 Len=0 MSS=1460 SACK_PERM=1
29416 18.018212 192.168.1.2 192.168.1.63 TCP iscsi-target > cap [ACK] Seq=1
Ack=2229090051 Win=65535 Len=0
29417 18.018318 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [RST]
Seq=2229090051 Win=0 Len=0
29418 18.439915 QuantaCo_9b:e4:69 Broadcast ARP Gratuitous ARP for
192.168.1.63 (Request)
29419 19.439930 QuantaCo_9b:e4:69 Broadcast ARP Gratuitous ARP for
192.168.1.63 (Request)
29420 20.986835 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [SYN] Seq=0
Win=65535 Len=0 MSS=1460 SACK_PERM=1
29421 20.986862 192.168.1.2 192.168.1.63 TCP [TCP Previous segment lost]
iscsi-target > cap [SYN, ACK] Seq=590980907 Ack=1 Win=14600 Len=0 MSS=1460
SACK_PERM=1
29422 20.986966 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [ACK] Seq=1
Ack=590980908 Win=65535 Len=0
29423 20.987203 192.168.1.63 192.168.1.2 iSCSI Login Command
29424 20.987222 192.168.1.2 192.168.1.63 TCP iscsi-target > cap [ACK]
Seq=590980908 Ack=197 Win=15544 Len=0
29422 20.986966 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [ACK] Seq=1
Ack=590980908 Win=65535 Len=0
29423 20.987203 192.168.1.63 192.168.1.2 iSCSI Login Command
29424 20.987222 192.168.1.2 192.168.1.63 TCP iscsi-target > cap [ACK]
Seq=590980908 Ack=197 Win=15544 Len=0
29425 20.987376 192.168.1.2 192.168.1.63 iSCSI Login Response (Success)
29426 20.987715 192.168.1.63 192.168.1.2 iSCSI Login Command
29427 20.987815 192.168.1.2 192.168.1.63 iSCSI Login Response (Success)
29428 21.190002 192.168.1.2 192.168.1.63 iSCSI [TCP Retransmission] Login
Response (Success)
29429 21.190154 192.168.1.63 192.168.1.2 TCP cap > iscsi-target [ACK]
Seq=545 Ack=590981300 Win=65143 Len=0
29430 30.987718 192.168.1.63 192.168.1.2 iSCSI SCSI: Report LUNs LUN: 0x00
29431 30.987878 192.168.1.2 192.168.1.63 iSCSI SCSI: Data In LUN: 0x00
(Report LUNs Response Data) SCSI: Response LUN: 0x00 (Report LUNs) (Good)
29432 30.988292 192.168.1.63 192.168.1.2 iSCSI SCSI: Inquiry LUN: 0x00

Could you give me some advise about what could be wrong? Is it possible that
iPXE can boot faster than local HDD? Thanks!

Regards,
Jerry

_______________________________________________
ipxe-devel mailing list
ipxe-devel at lists.ipxe.org
https://lists.ipxe.org/mailman/listinfo/ipxe-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ipxe.org/pipermail/ipxe-devel/attachments/20110522/53044189/attachment.htm>