[EXTERNAL EMAIL] 

TBH perccli/storcli are much better tools in my opinion, and if you're
looking for structured data, both will generate JSON output (using a final
command line argument of J, e.g., storcli /c0/eall/sall show J) and if
you're pushing measurements into something like Graphite or InfluxDB, it's
much easier to parse with something like Python than finessing strings with
grep. Just my $0.02, though! When I found storcli I never looked back (it
also supports megacli completely as well).

cheers,
Klaus

On Fri, Oct 11, 2019 at 7:16 AM Blake Hudson <[email protected]> wrote:

> [EXTERNAL EMAIL]
> Onno, I'm not sure how Dell allows you to configure the server at order
> time, but Dell often have configuration limitations that do not seem
> immediately obvious. As a concrete example, if one tries to configure a
> server with too much RAM or too many disks you may receive a warning that
> you have to upgrade the PSU in the server to complete your order.
>
> I could imagine that, as another possible case, Dell may allow you to
> install rear disks but it's possible they limit these disks to 7.2k (or
> disks rated to run at higher temperature, draw less power, etc). If you are
> installing disks yourself, you may not be aware of some of these
> limitations (as they may not be documented clearly or publicly). Every disk
> manufacturer sets their own temperature thresholds so while one drive may
> support 60 C, another may top out at 50 C. This isn't to say that either
> disk will be reliable if kept under those temperatures, just that those are
> the manufacturer's recommended operating temperatures and that they vary
> from model to model. If you're having high failure rates on rear disks, and
> the only obvious difference between front and rear disks is the operating
> temperature, I think that's a strong indicator that temperature could be a
> factor. Going forward you might consider using SSDs (which often produce
> less heat), lower rpm disks (that produce less heat), or disks rated for
> higher temperature extremes to see if there is a reliability improvement.
>
> Onno Zweers wrote on 10/11/2019 2:43 AM:
>
> Following up.
>
> I checked two classes of servers:
> R730xd - rear disk 58°C
> R740xd2 - rear disk 33°C
>
> That's a huge difference. The fan speeds were similar, between 10,300 and 
> 11,160 rpm. I don't think this accounts for the difference in temperature. 
> Perhaps the airflow of the system has been improved in the R740. But there is 
> one significant difference: in the R740xd2, the rear disks are SSDs, where 
> the R730xd have spinning disks.
>
> Cheers,
> Onno
>
>
> Op 10 okt. 2019, om 20:19 heeft Onno Zweers <[email protected]> 
> <[email protected]> het volgende geschreven:
>
> Thanks everyone for the very useful answers. I had a quick look:
>
> [root@shark5 ~]# for disk in $(smartctl --scan | egrep -o megaraid,[0-9]+) ; 
> do echo -n "$disk - " ; smartctl -a /dev/sdb -d $disk | grep 'Current Drive 
> Temperature' ; done
> megaraid,0 - Current Drive Temperature:     31 C
> megaraid,1 - Current Drive Temperature:     32 C
> megaraid,2 - Current Drive Temperature:     32 C
> megaraid,3 - Current Drive Temperature:     31 C
> megaraid,4 - Current Drive Temperature:     32 C
> megaraid,5 - Current Drive Temperature:     30 C
> megaraid,6 - Current Drive Temperature:     32 C
> megaraid,7 - Current Drive Temperature:     32 C
> megaraid,8 - Current Drive Temperature:     31 C
> megaraid,9 - Current Drive Temperature:     32 C
> megaraid,10 - Current Drive Temperature:     34 C
> megaraid,11 - Current Drive Temperature:     32 C
> megaraid,12 - Current Drive Temperature:     56 C
> megaraid,13 - Current Drive Temperature:     58 C
> megaraid,14 - Current Drive Temperature:     44 C
> megaraid,15 - Current Drive Temperature:     45 C
> megaraid,16 - Current Drive Temperature:     47 C
> megaraid,17 - Current Drive Temperature:     51 C
>
> 58 degrees C seems very hot to me, and indeed disks 12 and 13 are in the back 
> of the machine. We have lots of these servers and we've noticed that these 
> rear disks fail rather often. The 2 disks in the rear have as many failures 
> as the 12 disks in front. I guess the next step would be to check at which 
> speed the fans blowing.
>
> Cheers,
> Onno
>
> _______________________________________________
> Linux-PowerEdge mailing 
> [email protected]https://lists.us.dell.com/mailman/listinfo/linux-poweredge
>
>
> _______________________________________________
> Linux-PowerEdge mailing 
> [email protected]https://lists.us.dell.com/mailman/listinfo/linux-poweredge
>
>
> _______________________________________________
> Linux-PowerEdge mailing list
> [email protected]
> https://lists.us.dell.com/mailman/listinfo/linux-poweredge
>
_______________________________________________
Linux-PowerEdge mailing list
[email protected]
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

Reply via email to