Unsolved
This post is more than 5 years old
1 Rookie
•
69 Posts
0
2009
August 7th, 2017 08:00
VNX5300 & double fault ?
Hello,
I've a VNX5300 equipped with 50x HDDs (600GB each).
Bus0 Enclosure0 has:
- d0, d1, d2, d3: reserved
- d4: raidgroup 100 hot spare
- d5..d13: raidgroup 1 (9x hdd)
- d14..d22: raidgroup 2 (9x hdd)
- d23..d24: raidgroup 3 (2x hdd)
Bus1 Enclosure0 has:
- d0..d6: raidgroup 3 (7x hdd)
- d7..d15: raidgroup 4 (9x hdd)
- d16..d24: raidgroup 5 (9x hdd)
So I have a total of 5 * RAID5 (8+1) which as capacity totals of: 5 * (500*8) = 4 * 4000 = 20TB which match my current capacity.
Now the question I've for you is the following... I'm experiencing a bad HDD in B0E0D10 (raidgroup1) and the system is currently using d4 as hot spare. I was expecting instead that the faulted drive was taken care by its RAID5 first and the d4 in case of double fault.
Can anyone share some more light on the fault logic used by this VNX5300 ?
Thank you,
A.
No Events found!
Hendrajaya
13 Posts
0
August 8th, 2017 00:00
please refer to the following EMC community discussion, it explain very details how the hot-spare works.
VNX5300 Hot Spare
in summary: any available hotspare will automatically replaces the bad drives, double failure managed by raid if you have another disk failed on another raid group but you do not have any other hotspare left.
abuzzi
1 Rookie
•
69 Posts
0
August 8th, 2017 01:00
Hello,
thank you for the detailed hot-spare explanation link.
Just need a clarification about your summary:
"in summary: any available hotspare will automatically replaces the bad drives, double failure managed by raid if you have another disk failed on another raid group but you do not have any other hotspare left."
You wrote that double failure is now managed within raid since there are now no hotspares left.
But why you mention "another disk failed on another raid group" ? Would raidgroup1 still cope with an additional failure ? After all raidgroup1 is still composed by 8+1 disks though d4 replaced d10, isn't ?
Thank you,
A.
Hendrajaya
13 Posts
0
August 8th, 2017 06:00
it depend on the failed scenario (i am pointing to the scenario #1)
1. if you have 2 disk failure concurrently or when the spare disk have not fully synched, the RAID-5 will failed.
2. but if your spare has fully synched and there is another disk failed in the same RAID-5, the raid will survive
abuzzi
1 Rookie
•
69 Posts
0
August 8th, 2017 06:00
Thank you Hendrajaya !