[pca] patch issue

[Don O'Malley]

Hi,

Just to give you a quick update on this.

AFAICT we are expecting patches to be submitted to resolve CR 6797442 
(fp_cache_constructor()  causes fm-capability ereports) later today.
This would mean that, assuming that no issues are discovered during test 
& verification, these patches could be available as early the next 2 weeks.
The patch ids containing the fix will be 141874-05 & 141875-05.
If you have escalated this issue through the regular support channels 
you may be able to get access to the patch while it is in the T-Patch 
state (though this obviously comes with the caveat that the patch has 
not yet completed testing and should be used with caution).

I have also flagged the issue with our Sun Alert team and they are 
preparing a SunAlert for this issue.

There is another possible workaround being discussed to prevent the 
constant error notification generation:
Add the following line to /kernel/drv/emlxs.conf:
fm-capable=0;
Then reboot the host.
This workaround has one side effect; it will disable all DDI FMA related 
ereport, not just for "dvr-name = fp".
This is currently just a suggestion and has not yet been tested (also 
you will need to reboot), so I am not advocating you use it. I'm just 
sharing the information with you...

I'll pass on more information as I get it.

Best,
-Don

Dominique Frise wrote:
> Hi Don,
>
> We have about 10 Sparc servers with emlxs drivers impacted by this 
> issue. In other words: we are also waiting for a fix !
>
> Thanks,
>
> Dominique
>
>
> Don O'Malley wrote:
>> Hi Kevin/Brian,
>>
>> As this CR has Escalations open against it there should be a patch 
>> containing the fix available for it before May 2010.
>>
>> I'm not sure roughly when you could expect a fix; I'll do some 
>> digging and see if I can find out.
>>
>> Best,
>> -Don
>>
>> Bliss, Kevin L wrote:
>>> I spaced that one, but that makes the delivery date way out there, I 
>>> hope the SE is wrong or I will have to back out the existing patch. 
>>> Thanks
>>>
>>> -----Original Message-----
>>> From: pca-bounces at lists.univie.ac.at 
>>> [mailto:pca-bounces at lists.univie.ac.at] On Behalf Of Brian R Myers
>>> Sent: Tuesday, November 03, 2009 11:29 AM
>>> To: PCA (Patch Check Advanced) Discussion
>>> Subject: Re: [pca] patch issue
>>>
>>> Kevin,
>>>
>>> Update 8 was just released the beginning of last month.  If Sun's 
>>> schedule holds constant, update 9 should be released in May 2010.
>>>
>>> -Brian
>>>
>>>
>>> -----Original Message-----
>>> From: pca-bounces at lists.univie.ac.at 
>>> [mailto:pca-bounces at lists.univie.ac.at] On Behalf Of Bliss, Kevin L
>>> Sent: Tuesday, November 03, 2009 10:49 AM
>>> To: 'PCA (Patch Check Advanced) Discussion'
>>> Subject: Re: [pca] patch issue
>>>
>>> I opened a case with Sun, the engineer opened a bug, they found that 
>>> the bug was a duplicate so closed the new one.  The fix is 
>>> supposedly complete but won't be available until Sol 10 update 9 
>>> (which I thought was available).  In case you are interested the bug 
>>> is CR6797442.  Not as much detail as you received in your ticket but 
>>> neither is complete. 
>>> -----Original Message-----
>>> From: pca-bounces at lists.univie.ac.at 
>>> [mailto:pca-bounces at lists.univie.ac.at] On Behalf Of Hannu Kivimäki
>>> Sent: Tuesday, November 03, 2009 6:23 AM
>>> To: pca at lists.univie.ac.at
>>> Subject: Re: [pca] patch issue
>>>
>>> Hi,
>>>
>>> On Thu, 2009-10-22 at 11:09 -0700, Bliss, Kevin L wrote:
>>>  
>>>> This is not a pca issue but may be helpful to those on the list.  A
>>>> recent Solaris 10 patch has created a "compatibility" issue with fmd
>>>> which causes it to log 10+ errors per second to /var/fm/fmd/errlog.  I
>>>> only noticed because one of the systems patched was running low on
>>>> space.  I opened a case with Sun yesterday morning but so far have not
>>>> received any useful information, including which patch caused the
>>>> problem.  The system is the following error being logged several 
>>>> times a
>>>> second:
>>>>
>>>> Oct 21 2009 12:35:09.061332556 ereport.io.ddi.fm-capability
>>>> nvlist version: 0
>>>>         class = ereport.io.ddi.fm-capability
>>>>         ena = 0x2196595e6b07001
>>>>         detector = (embedded nvlist)
>>>>         nvlist version: 0
>>>>                 version = 0x0
>>>>                 scheme = dev
>>>>                 device-path = /
>>>>         (end detector)
>>>>
>>>>         dvr-name = fp
>>>>         __ttl = 0x1
>>>>         __tod = 0x4adf626d 0x3a7dc4c
>>>>     
>>>
>>> A quick note about this older topic - I patched some servers yesterday
>>> and
>>> noticed /var/fm/fmd/errlog getting filled with
>>> ereport.io.ddi.fm-capability
>>> messages (500+ MB of them overnight). Contacted Sun Support and they
>>> told
>>> me to back out Emulex patch 141876-05. FMA support was introduced in
>>> 141876-03 and apparently something broke it already in 141876-05.
>>>
>>> We never had any revisions of 141876 installed before, so after
>>> removing 141876-05, we are now back to Emulex patch 139608-05 with
>>> no fmd issues. I guess we have to wait for 141876-06 before trying
>>> patching Emulex HBA drivers again.
>>>
>>>
>>> Hannu
>>>
>>>
>>>
>>>
>>>
>>>
>>>   
>