<div dir="ltr"><div>I modified the patch to include whether it's IPv6 or IPv4 IOCtl that's failed:</div><div><br></div><div> ret = ioctl(s, SIOCGIFCONF, ifc);<br> if (ret < 0) {<br> acm_log(0, "ioctl IPv%s ifconf error: %s\n",<br> (family == AF_INET6) ? "6" : "4", strerror(errno));<br> goto out2;<br> }<br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Oct 25, 2017 at 10:55 AM, Michael Di Domenico <span dir="ltr"><<a href="mailto:mdidomenico4@gmail.com" target="_blank">mdidomenico4@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">while i totally agree with both you and Hal and the fact that this is<br>
probably a rare issue. my contention really stems from the point that<br>
the error message provides no clue that it's an ipv6 error<br>
<div class="HOEnZb"><div class="h5"><br>
On Wed, Oct 25, 2017 at 10:46 AM, Weiny, Ira <<a href="mailto:ira.weiny@intel.com">ira.weiny@intel.com</a>> wrote:<br>
> Agreed. I don’t know of many people who completely disable IPv6. So this<br>
> should be rare. And if they do then they should know that they will get<br>
> AF_INET6 errors on any software which is trying to support both…<br>
><br>
><br>
><br>
> From: Hal Rosenstock [mailto:<a href="mailto:hal.rosenstock@gmail.com">hal.rosenstock@gmail.<wbr>com</a>]<br>
> Sent: Wednesday, October 25, 2017 5:56 AM<br>
> To: Michael Di Domenico <<a href="mailto:mdidomenico4@gmail.com">mdidomenico4@gmail.com</a>><br>
> Cc: Weiny, Ira <<a href="mailto:ira.weiny@intel.com">ira.weiny@intel.com</a>>; <a href="mailto:users@lists.openfabrics.org">users@lists.openfabrics.org</a><br>
> Subject: Re: [Users] ibacm?<br>
><br>
><br>
><br>
> Right now, I don't see obvious way to eliminate that message for AF_INET6 as<br>
> sometimes it's appropriate and other times not.<br>
><br>
><br>
><br>
> On Wed, Oct 25, 2017 at 8:44 AM, Michael Di Domenico<br>
> <<a href="mailto:mdidomenico4@gmail.com">mdidomenico4@gmail.com</a>> wrote:<br>
><br>
> the patch does work. the only complaint i might register is that the<br>
> first time through the loop you still get an ifconf ioctl error in the<br>
> log file. it does proceed along to bind to the AF_INET afterwards,<br>
> but it might cause a red herring support issue later on<br>
><br>
><br>
> On Tue, Oct 24, 2017 at 5:30 PM, Hal Rosenstock<br>
> <<a href="mailto:hal.rosenstock@gmail.com">hal.rosenstock@gmail.com</a>> wrote:<br>
>> I have supplied a patch for Michael to try and will submit it to rdma-core<br>
>> once tested...<br>
>><br>
>> On Tue, Oct 24, 2017 at 4:38 PM, Weiny, Ira <<a href="mailto:ira.weiny@intel.com">ira.weiny@intel.com</a>> wrote:<br>
>>><br>
>>> We probably need to fall back to an AF_INET check. I did not realize<br>
>>> that<br>
>>> disabling ipv6 would cause this to fail because AF_INET6 usually “covers”<br>
>>> AF_INET. My guess here is that if you excluded IPv6 support from the<br>
>>> kernel<br>
>>> that would explain the failure.<br>
>>><br>
>>><br>
>>><br>
>>> Perhaps you could try a patch which falls back to AF_INET?<br>
>>><br>
>>><br>
>>><br>
>>> Ira<br>
>>><br>
>>><br>
>>><br>
>>> From: Users [mailto:<a href="mailto:users-bounces@lists.openfabrics.org">users-bounces@lists.<wbr>openfabrics.org</a>] On Behalf Of Hal<br>
>>> Rosenstock<br>
>>> Sent: Tuesday, October 24, 2017 8:59 AM<br>
>>> To: Michael Di Domenico <<a href="mailto:mdidomenico4@gmail.com">mdidomenico4@gmail.com</a>><br>
>>> Cc: <a href="mailto:users@lists.openfabrics.org">users@lists.openfabrics.org</a><br>
>>> Subject: Re: [Users] ibacm?<br>
>>><br>
>>><br>
>>><br>
>>> This makes more sense ;-)<br>
>>><br>
>>><br>
>>><br>
>>> What were the steps used to disable ipv6 ?<br>
>>><br>
>>><br>
>>><br>
>>> On Tue, Oct 24, 2017 at 11:16 AM, Michael Di Domenico<br>
>>> <<a href="mailto:mdidomenico4@gmail.com">mdidomenico4@gmail.com</a>> wrote:<br>
>>><br>
>>> it looks like i found the culprit<br>
>>><br>
>>> when running ibacm out of the box on rhel 7.4 the ibacm.log shows<br>
>>><br>
>>> acm_if_iter_sys: ioctl ifconf error -1<br>
>>><br>
>>> if i change the line<br>
>>><br>
>>> s = socket(AF_INET6, SOCK_DGRAM, 0);<br>
>>><br>
>>> to<br>
>>><br>
>>> s = socket(AF_INET, SOCK_DGRAM, 0);<br>
>>><br>
>>> as Hal suggested and start the ibacm daemon i acm correctly bind to be<br>
>>> ipoib addresses and interfaces. my initial report of this change not<br>
>>> being effective was a miss communication in that i though it related<br>
>>> to the client and not the service process<br>
>>><br>
>>> i can then run<br>
>>><br>
>>> ib_acme -f i -s 172.22.64.96 -d 172.22.64.96 -S 172.22.64.96 -v V<br>
>>><br>
>>> and get back valid data<br>
>>><br>
>>> ib_acme -d <hostname> still doesn't work, but that might be internal<br>
>>> we don't currently have reverse/forward entries for our ipoib<br>
>>> interfaces, i'm still looking into it<br>
>>><br>
>>><br>
>>><br>
>>><br>
>>> On Mon, Oct 16, 2017 at 1:54 PM, Michael Di Domenico<br>
>>> <<a href="mailto:mdidomenico4@gmail.com">mdidomenico4@gmail.com</a>> wrote:<br>
>>> > it's come to my attention that ibacm might not be working correctly on<br>
>>> > my cluster, but i'm unable to determine why ibacm is failing<br>
>>> ><br>
>>> > here's what i did<br>
>>> ><br>
>>> > ib_acme -A -O<br>
>>> > systemctl restart ibacm<br>
>>> ><br>
>>> > in the /var/log/ibacm file i see<br>
>>> ><br>
>>> > acm_if_iter_sys: ioctl ifconf error -1<br>
>>> > acmp_join_group: qib0 1 pkey 0xffff, sl 0x0, rate 0x3, mtu 0x4<br>
>>> > acm_server: started<br>
>>> ><br>
>>> > but when i try to query the node locally<br>
>>> ><br>
>>> > ib_acme -d node001 -v -V<br>
>>> ><br>
>>> > in the log file i see<br>
>>> ><br>
>>> > acm_svr_resolve_dest: notice - unknown local end point address<br>
>>> ><br>
>>> > on the console i see<br>
>>> ><br>
>>> > Service: localhost<br>
>>> > Destination: 172.21.80.1<br>
>>> > ib_acm_resolve_ip failed: cannot assign requested address<br>
>>> > SA verification: failed cannot assign requested address<br>
>>> ><br>
>>> > the ibacm_addr.cfg contains<br>
>>> > node001 qib0 1 default<br>
>>> > node001-1 qib0 1 default<br>
>>> ><br>
>>> > all the nodes in the cluster are configured the exact same way. and<br>
>>> > produce the same result when trying to query locally or a remote node<br>
>>> ><br>
>>> > any thoughts?<br>
>>> ______________________________<wbr>_________________<br>
>>> Users mailing list<br>
>>> <a href="mailto:Users@lists.openfabrics.org">Users@lists.openfabrics.org</a><br>
>>> <a href="http://lists.openfabrics.org/mailman/listinfo/users" target="_blank" rel="noreferrer">http://lists.openfabrics.org/<wbr>mailman/listinfo/users</a><br>
>>><br>
>>><br>
>><br>
>><br>
><br>
><br>
</div></div></blockquote></div><br></div>