Hi Alistair, all,
Sorry for using three emails instead of one (which is exactly what you
were grumbling about!)
In summary we've patched all our worker nodes, SL6 and Centos7, all our
CEs, mix of SL6 and Centos7, the service nodes ARGUS, BDII, APEL, and
the storage systems. Middleware used here is ARC/Condor, VAC, DPM; so
far, no ill-effects have been encountered (touch wood, fingers crossed
etc.) We've also done some benchmarking (what else can we measure
easily?) and the hit is~ 2% in terms of HS06.
Unfortunately, the microcode_ctl package only covers some of our
systems, hence much of our cluster is still vulnerable to Variant 2.
https://security.web.cern.ch/security/advisories/spectre-meltdown/spectre-meltdown.shtml
We expect new microcode from Intel and the distros. I fear another
reboot will be needed on some/most of our nodes when that shows up in
the repos.
Cheers,
Ste
On 2018-01-10 09:17, sjones wrote:
> Also:
>
> DPM Head, DPM Pool
>
> Patched. Result: "Works for us".
>
> Ste
>
> On 2018-01-10 09:14, sjones wrote:
>> Hi Alistair,
>>
>> On 2018-01-10 00:22, Alastair Dewhurst wrote:
>>> Anyone done an ARC CE (+ HTCondor Schedd)? They aren’t massively
>>> loaded but they are constantly busy machines and could potentially be
>>> impacted by performance slow downs. Our intention is to do one on
>>> Wednesday and see if anything happens before doing the rest over the
>>> next few days.
>>
>>> ARC CE (+ HTCondor Schedd), BDII, Argus
>>
>> Yes. Result: "Works for me".
>>
>>> MyProxy Squids, PerfSonar
>>
>> Not sure, more work needed...
>>
>> Cheers,
>>
>> Ste
>> Liverpool T2
|