<html><body>
<p><br>
Wen Hao Wang (王文昊)<br>
<br>
Software Engineer<br>
IBM China Software Development Laboratory<br>
Email: wangwhao@cn.ibm.com<br>
Tel: 86-10-82451055<br>
Fax: 86-10-82782244 ext. 2312<br>
Address: 1/F, IBM ZGC Campus. Ring Building 28,ZhongGuanCun Software Park,No.8 Dong Bei Wang West Road, Haidian District Beijing, 100193, P.R.China<br>
<br>
<br>
<tt>Doug Ledford <dledford@redhat.com> 写于 2009-02-14 00:13:32:<br>
<br>
> On Fri, 2009-02-13 at 08:05 +0800, Wen Hao Wang wrote:<br>
> > Doug Ledford <dledford@redhat.com> 写于 2009-02-12 21:20:30:<br>
> > <br>
> > > On Thu, 2009-02-12 at 13:20 +0200, Tziporet Koren wrote:<br>
> > > > Wen Hao Wang wrote:<br>
> > > > ><br>
> > > > > Hi all:<br>
> > > > ><br>
> > > > > I changed my blade OS to RHEL5.3 yesterday and installed OFED<br>
> > (shipped <br>
> > > > > in RHEL5.3 image) by "yum groupisntall". Then I load some<br>
> > drivers and <br>
> > > > > wrote network interface configuration file ifcfg-ib0. ifup ib0<br>
> > also <br>
> > > > > succeeded. But IB utilites report Connetion timed out.<br>
> > > > ><br>
> > > > ><br>
> > > > > [root@xblade06 network-scripts]# sminfo<br>
> > > > > ibwarn: [32593] _do_madrpc: recv failed: Connection timed out<br>
> > > > > ibwarn: [32593] mad_rpc: _do_madrpc failed; dport (Lid 9)<br>
> > > > > sminfo: iberror: failed: query<br>
> > > > ><br>
> > > > > I had to reboot the blade and rerun "openibd start". Then<br>
> > sminfo <br>
> > > > > reported correct contents. I do not suppose this reboot is<br>
> > required. <br>
> > > > > Did I miss any configuration step?<br>
> > > <br>
> > > There was an unintentional bug in the rhel5.2 openibd init script in<br>
> > > that it automatically turned itself on during install (generally,<br>
> > most<br>
> > > init scripts should default to *not* turning themselves on during<br>
> > > install of the package, nor should they start themselves during<br>
> > install<br>
> > > of the package...this is for security reasons, imagine if you<br>
> > installed<br>
> > > the bind name server on your box and it automatically started up<br>
> > before<br>
> > > you had a chance to configure it). In rhel5.3 we fixed that bug.<br>
> > So,<br>
> > <br>
> > Yeah. I heard of this bug.<br>
> > <br>
> > > you may need to 'chkconfig --level 2345 openibd on' to make sure<br>
> > openibd<br>
> > > starts up each time. The error you list above is consistent with<br>
> > not<br>
> > > all of the kernel modules being loaded when you tried to use the<br>
> > sminfo<br>
> > > program.<br>
> > <br>
> > Even after reboot, service openibd is not started automatically.<br>
> > [root@xblade06 ~]# chkconfig --list openibd<br>
> > openibd 0:off 1:off 2:off 3:off 4:off 5:off 6:off<br>
> <br>
> That's because you have to run the command I listed in my first email to<br>
> turn it on.<br>
></tt><br>
<br>
<tt>I totally agree with this. But I am still confused why sminfo gave errors</tt><br>
<tt>before reboot, or which steps I should take for the first OFED usage before</tt><br>
<tt>reboot. As far as I can see, whether the service is added into system</tt><br>
<tt>runlevel DB is not related to the sminfo error. Please correct me if that</tt><br>
<tt>is not the case.</tt><br>
<tt><br>
> > I agree with you that maybe some modules were not loaded. But what's<br>
> > that?<br>
> > Before reboot, I run "/etc/init.d/openibd start" and<br>
> > "/etc/init.d/network<br>
> > restart". No error was reported. "openibd status" also looked good.<br>
> <br>
> Running start on a service does not enable that service at the next<br>
> reboot. You must specifically enable the service in order for it to<br>
> start automatically.<br>
> <br>
> > > <br>
> > > > > Moreover, "openibd start" report one warning message about<br>
> > hwconf. <br>
> > > > > Anyone has comments about this?<br>
> > > > ><br>
> > > > > [root@xblade07 ~]# /etc/init.d/openibd start<br>
> > > > > Loading OpenIB kernel modules:grep: /etc/sysconfig/hwconf: No<br>
> > such <br>
> > > > > file or directory<br>
> > > > > [ OK ]<br>
> > > <br>
> > > Can you see if the kudzu package is installed on your machine? The<br>
> > > openib package uses this config file written by kudzu to determine<br>
> > what<br>
> > > hardware drivers to load. I suppose I should put a specific<br>
> > requires in<br>
> > > the rpm for that.<br>
> > <br>
> > kudzu is installed.<br>
> > [root@xblade06 ~]# rpm -q kudzu<br>
> > kudzu-1.2.57.1.21-1<br>
> <br>
> Make sure kudzu has been run at least once then (it would appear to be<br>
> turned off on your machine or else /etc/sysconfig/hwconf would exist).<br>
> You can run it manually from the command line and that should be<br>
> sufficient for the openibd init script's needs.<br>
> </tt><br>
<br>
<tt>Yes. After kudza created the file on my machine, openibd script had no error</tt><br>
<tt>this time. I want to know in my scenario, is "openibd restart" needed/required?</tt><br>
<br>
<tt>Many thanks!</tt><br>
<br>
Wen Hao Wang<br>
Email: wangwhao@cn.ibm.com<br>
<tt><br>
> -- <br>
> Doug Ledford <dledford@redhat.com><br>
> GPG KeyID: CFBFF194<br>
> <a href="http://people.redhat.com/dledford">http://people.redhat.com/dledford</a><br>
> <br>
> Infiniband specific RPMs available at<br>
> <a href="http://people.redhat.com/dledford/Infiniband">http://people.redhat.com/dledford/Infiniband</a><br>
> <br>
> [附件 "signature.asc" 被 Wen Hao Wang/China/IBM 删除]</tt></body></html>