Hi Ted,
I'm quite sure I didn't observe any of these timeout/errors in
snv_75. Looking at the network pings from our network monitor graphs,
all the outages occur after I liveupgraded to snv_95. It's a fairly
typical SAMP system... I'm not running anything specific to trigger the
network outages. Some of the apps that are running are: Apache2, PHP 5,
MySQL, Mailman, Dovecot, Postfix, etc. The systemboard is a Supermicro
5015M-MT+ with the dual Intel e1000g NICs.
cheers,
steve
Ted You wrote:
> Hi Stephen,
>
> From the ereports, it seems that the device got hang and was reset by
> the driver automatically. From snv_77, the e1000g driver started to
> support FMA, so we see the FMA ereports now. The problem might have
> existed in snv_75.
>
> We have not seen this problem before. The recent bug fixes in snv_95
> have nothing to do with this problem. I need to try to reproduce the
> problem on our local systems. Could you please let me know what
> applications or tests you have been running on your system?
>
> Thanks,
> Ted
>
>
> Stephen Lau :
>> I'm seeing flaky outages on my snv_95 system (just recently LU'd from
>> snv_75 where I didn't have any issues).
>>
>> I seem to get outages of a few minutes (5-6) throughout the day, with
>> no rhyme or reason as to why.
>>
>> fmdump -e shows repeated occurrences of:
>> Aug 18 12:24:57.2612 ereport.io.device.stall Aug 18
>> 12:24:57.5985 ereport.io.service.restored
>> /usr/X11/bin/scanpci shows my e1000g devices to be:
>> pci bus 0x000d cardnum 0x00 function 0x00: vendor 0x8086 device 0x108c
>> Intel Corporation 82573E Gigabit Ethernet Controller (Copper)
>>
>> pci bus 0x000e cardnum 0x00 function 0x00: vendor 0x8086 device 0x109a
>> Intel Corporation 82573L Gigabit Ethernet Controller
>>
>>> From my reading it looks like there were 1 or 2 e1000g issues that
>>> were purportedly fixed in snv_95, but I'm still seeing these
>>> problems. Anyone have any idea as to what might be the issue, and
>>> whether or not there is a workaround? I'm debating about pulling
>>> back the e1000g driver from my snv_75 LU slice and seeing if that
>>> works, but I'm hoping someone has a workaround for now. :)
>>
>>
>> This message posted from opensolaris.org
>> _______________________________________________
>> networking-discuss mailing list
>> [email protected]
--
stephen lau | [EMAIL PROTECTED] | www.whacked.net
_______________________________________________
networking-discuss mailing list
[email protected]
