[openib-general] Re: uDAPL again

Aniruddha Bohra bohra at cs.rutgers.edu
Wed Nov 2 11:16:12 PST 2005


James Lentini wrote:

>On Wed, 2 Nov 2005, Aniruddha Bohra wrote:
>
>  
>
>>James Lentini wrote:
>>
>>    
>>
>>>On Wed, 2 Nov 2005, Aniruddha Bohra wrote:
>>>
>>> 
>>>      
>>>
>>>>Hello,
>>>>  The following is the log for a request I am sending,
>>>>
>>>>The number of IOVs for req is 2. And the iov is shown below :
>>>>
>>>>REQ[0] = (0xb5f3f100, 48, 0xca88003b)^M
>>>>REQ[1] = (0xb5f3f2b8, 152, 0xca88003b)^M
>>>>
>>>>dapl_ep_post_send (0x8087110, 2, 0x808b300, 0xb5f3f6b4, 0)^M
>>>>dapl_ep_post_send : LOCALIOV[0] = (0xb5f3f100, 48, 0xca88003b)^M
>>>>dapl_ep_post_send : LOCALIOV[1] = (0xb5f3f2b8, 152, 0xca88003b)^M
>>>>post_snd: ep 0x8087110 op 2 ck 0x8087374 sgs 2 l_iov 0x808b300 r_iov
>>>>0xbf964290 f 0^M
>>>>post_snd: ep 0x8087110 cookie 0x8087374 segs 2 l_iov 0x808b300^M
>>>>post_snd_localiov: lkey 0xca88003b va 0xb5f3f100 len 48 ^M
>>>>post_snd: lkey 0xca88003b va 0xb5f3f100 len 48 ^M
>>>>post_snd_localiov: lkey 0xca88003b va 0xb5f3f2b8 len 152 ^M
>>>>post_snd: lkey 0xca88003b va 0xb5f3f2b8 len 152 ^M
>>>>post_snd: op 0x2 flags 0x2 sglist 0xbf9641b0, 2^M
>>>>post_snd: returned^M
>>>>dapl_ep_post_send () returns 0x0^M
>>>>dapl_evd_wait (0x8083ca0, -1, 1, 0xbf9642d0, 0xbf9642cc)^M
>>>>dapl_evd_wait: EVD 0x8083ca0, CQ 0x8083da0^M
>>>>cq_object_wait: CQ channel 0x8081290 time -1^M
>>>>cq_object_wait: RET evd 0x8083ca0 ibv_cq 0x8083da0 ibv_ctx (nil) Success^M
>>>>       >>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<^M
>>>>       dapl_evd_dto_callback : CQE ^M
>>>>               work_req_id 134771572^M
>>>>               status 12^M
>>>>   
>>>>        
>>>>
>>>Status 12 is IBV_WC_RETRY_EXC_ERR.
>>>
>>>Are you sure you can communicate over IB? Do pings over IPoIB work, etc.?
>>>
>>> 
>>>      
>>>
>>bohra at hora-3 ~]$ ping -b 10.10.10.255
>>WARNING: pinging broadcast address
>>PING 10.10.10.255 (10.10.10.255) 56(84) bytes of data.
>>64 bytes from 10.10.10.12: icmp_seq=0 ttl=64 time=0.034 ms
>>64 bytes from 10.10.10.13: icmp_seq=0 ttl=64 time=8.98 ms (DUP!)
>>64 bytes from 10.10.10.12: icmp_seq=1 ttl=64 time=0.033 ms
>>64 bytes from 10.10.10.13: icmp_seq=1 ttl=64 time=0.095 ms (DUP!)
>>64 bytes from 10.10.10.12: icmp_seq=2 ttl=64 time=0.025 ms
>>64 bytes from 10.10.10.13: icmp_seq=2 ttl=64 time=0.096 ms (DUP!)
>>
>>--- 10.10.10.255 ping statistics ---
>>    
>>
>
>
>I don't see DUPs when I ping the broadcast address. Is it possible 
>another machine is configured with the same IP address?
>
>  
>
No. There are 3 nodes on the switch. Two openib nodes.

I can login to the .13 node through the 10.10.10 subnet using IPoIB.

bohra at hora-3 ~]$ ping 10.10.10.13
PING 10.10.10.13 (10.10.10.13) 56(84) bytes of data.
64 bytes from 10.10.10.13: icmp_seq=0 ttl=64 time=0.112 ms
64 bytes from 10.10.10.13: icmp_seq=1 ttl=64 time=0.063 ms
64 bytes from 10.10.10.13: icmp_seq=2 ttl=64 time=0.073 ms
64 bytes from 10.10.10.13: icmp_seq=3 ttl=64 time=0.052 ms
64 bytes from 10.10.10.13: icmp_seq=4 ttl=64 time=0.069 ms
64 bytes from 10.10.10.13: icmp_seq=5 ttl=64 time=0.055 ms
64 bytes from 10.10.10.13: icmp_seq=6 ttl=64 time=0.076 ms
64 bytes from 10.10.10.13: icmp_seq=7 ttl=64 time=0.052 ms
64 bytes from 10.10.10.13: icmp_seq=8 ttl=64 time=0.070 ms



Aniruddha





More information about the general mailing list