Once upon a time, Gary Stainburn <gary@xxxxxxxxxxxxxx> said:
> I've used MDADM before on previous servers, but have found that this setup
> isn't hot swap. Ultimately if I had to replace a drive it involved a lot of
> effort, especially the first drive.
I use mdadm RAID in a bunch of places; it isn't automated, but it isn't
that hard to replace a drive. It would be nice if the storaged project
had support for this (I think the various needed bits are there, just
needs somebody to write a front-end to make the calls I guess).
> By using H/W RAID, it's literally just a case of removing the dead drive and
> inserting the replacement. I've got a number of IBM and DELL boxes like this.
> it's just a pity they're not compatible with Linux so I can't monitor or
> manage them while the servers are running. The only way I know I have
> problems is by watching the LEDS
I also use Dell servers with various hardware RAID cards in a bunch of
places. I install the Dell tools (they have yum repos for this), and
then can use omreport and omconfig to monitor and manage RAID from
within Linux, and their SNMP agent to monitor system health from my
external monitoring system.
One nice thing with omreport/omconfig is that it doesn't matter what
type of RAID card/chip/whatever they use (because they change things
over generations), the commands are the same.
Newer Dell servers also have integrated RAID management to the DRAC
directly, so you can monitor/manage/etc. through the out-of-band web and
SSH UI.
--
Chris Adams <linux@xxxxxxxxxxx>
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
https://lists.centos.org/mailman/listinfo/centos
