[Spread-users] Spread Retransmission Help
John Schultz
jschultz at spreadconcepts.com
Mon Oct 7 12:40:02 EDT 2013
I'd try longer runs with smaller bursts:
./spend -a 225.1.2.3 -b 10 -n 10000
If you keep seeing such high loss, then something seems to be wrong somewhere in your network. It could be a bad NIC, bad cable, bad port on a switch, misconfigured or malfunctioning router or switch. I'd try also try using point to point addresses and see if you have similar problems or not. Try all pairs and see if you can narrow down where exactly the problem lies. You can also try ping and see if you see similar problems.
Cheers!
-----
John Lane Schultz
Spread Concepts LLC
Cell: 443 838 2200
On Oct 7, 2013, at 12:22 PM, Lane Matthew wrote:
Thanks for the quick replies.
I have tested using spsend and sprecv as suggested. What I have found is that the number of 'missed' varies each time I run the send. Sometimes it is 0; other times it is 90%.
> spsend -a 225.1.2.3 -n 100
Checking (225.1.2.3, 4444). Each burst has 100 packets, 1024 bytes each with 10 msec delay in between, for a total of 100 packets
total time is (0,31000), with 0 problems
> sprecv -a 225.1.2.3 -d
-------
--> count is 30, i is 28, missed 2 total missed 2, corrupt 0
--> count is 73, i is 39, missed 34 total missed 36, corrupt 0
--> count is 94, i is 82, missed 12 total missed 48, corrupt 0
-------
Report: total packets 100, total missed 48, total corrupted 0
-------
Thanks
Matthew Lane
-----Original Message-----
From: John Schultz [mailto:jschultz at spreadconcepts.com]
Sent: 07 October 2013 16:27
To: Lane Matthew
Cc: spread-users at lists.spread.org
Subject: Re: [Spread-users] Spread Retransmission Help
Given that subnet mask, that is the proper broadcast address. If this network is more complicated than a single switch or router, then it might be that broadcast is turned off or not working for parts of it. You could also try using a multicast address rather than the broadcast address to see if that works better.
Are you sure there are no firewalls interfering with any of the ports (14803 UDP and 14804 UDP)?
As Yair suggested you could build and use the spsend and sprecv programs to try to find why your loss is so significant.
Cheers!
-----
John Lane Schultz
Spread Concepts LLC
Cell: 443 838 2200
On Oct 7, 2013, at 11:02 AM, Lane Matthew wrote:
Thanks.
I thought this too so I verified the broadcast address using the IP and subnet mask. Subnet Mask=255.255.254.0 for all servers. Looking online, 10.151.141.255 is the network broadcast address for this subnet. Have I misunderstood something in the setup?
Thanks
Matthew Lane
-----Original Message-----
From: John Schultz [mailto:jschultz at spreadconcepts.com]
Sent: 07 October 2013 15:49
To: Lane Matthew
Cc: spread-users at lists.spread.org
Subject: Re: [Spread-users] Spread Retransmission Help
I think your network addresses are wrong.
Your segment broadcast address is 10.151.141.255 while your machines are sitting in the 10.151.140 network. That probably isn't right, so your segment address probably isn't working at all and Spread is working hard to overcome that.
Cheers!
-----
John Lane Schultz
Spread Concepts LLC
Cell: 443 838 2200
On Oct 7, 2013, at 10:38 AM, Lane Matthew wrote:
Hi,
Whilst looking at the spread logs; I noticed verion high retans numbers. Is this expected or does it suggest I have a problem? If the latter; how would you suggest I could go about indentifying what the problem could be?
Config
Spread_Segment 10.151.141.255:14803 {
SERVER1 10.151.140.236
SERVER2 10.151.140.237
SERVER3 10.151.140.234
SERVER4 10.151.140.235
}
Spread Logs
============================
Status at SERVER4 V 4.00. 0 (state 1, gstate 1) after 4641117 seconds :
Membership : 4 procs in 1 segments, leader is SERVER1
rounds : 24279316 tok_hurry : 19921464 memb change: 21
sent pack: 24535749 recv pack : 77414263 retrans : 8423199
u retrans: 6694629 s retrans : 1728570 b retrans : 0
My_aru : 15433449 Aru : 15433449 Highest seq: 15433449
Sessions : 116 Groups : 1977 Window : 60
Deliver M: 57102298 Deliver Pk: 89528715 Pers Window: 15
Delta Mes: 1233 Delta Pack: 215 Delta sec : 25
==================================
Monitor>
============================
Status at SERVER3 V 4.00. 0 (state 1, gstate 1) after 2659181 seconds :
Membership : 4 procs in 1 segments, leader is SERVER1
rounds : 24279315 tok_hurry : 11045561 memb change: 14
sent pack: 13845399 recv pack : 40335734 retrans : 2957258
u retrans: 1117748 s retrans : 1839510 b retrans : 0
My_aru : 15433449 Aru : 15433449 Highest seq: 15433449
Sessions : 53 Groups : 1977 Window : 60
Deliver M: 29916378 Deliver Pk: 47580732 Pers Window: 15
Delta Mes: -27185920 Delta Pack: 0 Delta sec : -1981936
==================================
Monitor>
============================
Status at SERVER1 V 4.00. 0 (state 1, gstate 1) after 3822591 seconds :
Membership : 4 procs in 1 segments, leader is SERVER1
rounds : 24279316 tok_hurry : 18360366 memb change: 14
sent pack: 14094574 recv pack : 72830765 retrans : 6315107
u retrans: 3723218 s retrans : 2591889 b retrans : 0
My_aru : 15433449 Aru : 15433449 Highest seq: 15433449
Sessions : 81 Groups : 1977 Window : 60
Deliver M: 47437911 Deliver Pk: 74928112 Pers Window: 15
Delta Mes: 17521533 Delta Pack: 0 Delta sec : 1163410
==================================
Monitor>
============================
Status at SERVER2 V 4.00. 0 (state 1, gstate 1) after 4641102 seconds :
Membership : 4 procs in 1 segments, leader is SERVER1
rounds : 24279315 tok_hurry : 19920799 memb change: 16
sent pack: 16699063 recv pack : 81794705 retrans : 10967659
u retrans: 7755320 s retrans : 3212339 b retrans : 0
My_aru : 15433449 Aru : 15433449 Highest seq: 15433449
Sessions : 72 Groups : 1977 Window : 60
Deliver M: 57101208 Deliver Pk: 89523128 Pers Window: 15
Delta Mes: 9663297 Delta Pack: 0 Delta sec : 818511
==================================
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Message for the recipient only, if received in error, please notify the sender and read http://www.eni.com/disclaimer/ _______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users
_______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users
_______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users
More information about the Spread-users
mailing list