xfs_repair fails after trying to format log cycle?

[Andrew Ryder]

On 06/06/2016 03:19 PM, Andrew Ryder wrote:
>
>
> On 06/06/2016 11:33 AM, Emmanuel Florac wrote:
>> Le Fri, 3 Jun 2016 22:28:54 -0400
>> Andrew Ryder <tireman at shaw.ca> écrivait:
>>
>>> Sorry to dig up old stuff here but after replacing hardware and
>>> pulling my hair out to no end, I traced the source of the problem out.
>>>
>>> On every drive in the array, the points on the PCB which the contacts
>>> for the drives motors and head/actuators make contact with were
>>> oxidized enough to cause the issue. I ended up pulling the PCB off
>>> each drive and cleaning them as well as cleaning out all other cable
>>> and drive connectors from the HBA outward and everything is happy
>>> again.
>>>
>>> http://s33.postimg.org/uhjmvw4dr/Not_Cleaned.jpg
>>> http://s33.postimg.org/xo94ieii7/Partial_Cleaned_1.jpg
>>> http://s33.postimg.org/hoqgyumgf/Partial_Cleaned_2.jpg
>>> http://s33.postimg.org/68k20t8a7/Partial_Cleaned_3.jpg
>>>
>>
>> Good job, that was quite unexpected :) Are you next to the
>> seashore?
>>
>
>
> No. I'm up in NW Ontario in Canada actually, far away from any salt
> water. I'm guessing it was from the combination of air
> pollutants/humidity and heat given the drives are not in the most
> hospitable environment esp during the summer ... but after cleaning the
> board contacts up as well as the esata cable ends/ports, its working
> great again.
>
> I'm beginning to think when I do replace any drives, after warranty is
> up on them, pull off the PCB, clean it up and coat the contacts with a
> good smearing of dielectric grease to keep the issue from happening.
>
> Makes me wonder how many other people out there suffer from the same
> issue given the amount of unresolved similar hw errors I've found while
> googling the errors..
>
> Thanks again,
> Andrew
>

One other thing might be worth noting, before I cleaned up the oxidation 
on the drives boards, they would work fine if on dedicated SATA channels 
off the mainboard, but when put on any sata port multiplier, it would do 
nothing but act up. Which was really weird.

I'm assuming there must be some inherent sensitivity with sata PMP 
chips/hw that make this more prevalent.. and I've tried both the old and 
new version of SansDigital's TR8M's backplanes, 2 different sata PMP 
HBAs (rocketraid 2722 and addonics older PCI 4 port sata pmp hba) as 
well as an older addonics 3gbit sata port multiplier fan out board ..

> _______________________________________________
> xfs mailing list
> xfs at oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs