On 2/6/19 11:49 AM, Chris Murphy wrote:
On Wed, Feb 6, 2019 at 9:59 AM Kevin Fenzi <kevin(a)scrye.com>
wrote:
>
> On 2/6/19 7:23 AM, Vít Ondruch wrote:
>>
>> Dne 05. 02. 19 v 21:19 Kevin Fenzi napsal(a):
>>> On 2/5/19 3:50 AM, Peter Robinson wrote:
>>>>> There seems to be an issue with moving builds over. Everything seems
to be stuck in f30-pending and not getting tagged with f30.
>>>> That'll be the signing queue
>>> Yeah, it wasn't processing these correctly I'm afraid.
>>>
>>> I did add the mass rebuild side tag to autosign, but it was a day or two
>>> after it started so likely these are the ones before that.
>>>
>>> The autosign machine also had some issues the last few days (it lost 3
>>> disks in a row) and we did some reboots to replace things, etc.
>>>
>>> In any case, I have a loop running telling robosign to sign and move all
>>> those f30-pending ones. It's slowly processing them down. Sadly, I
don't
>>> think it will be done tonight, but we will see.
>>
>>
>> It does not look better, does it? What is the status please?
>
> Our autosigning machine seems to have a failed tpm in it, which has
> caused signing to fail to work. ;(
>
> We are working on getting hardware fixed, or a replacement in place.
>
> More updates as I know more. ;(
>
> This is really a anoying time for hardware to die, but of course
> sometimes it seems like it _knows_ when that is so it can fail.
Is it for sure dead, or is it possible it coincided with a kernel
update? I recently saw some messages of a TPM regression on the
linux-integrity@ list, but I'm not remembering the kernel versions
affected.
The affected machine was running the rhel7 kernel, and we did boot into
the kernel it's been using for a long time without any change. ;(
kevin