Spare space problem.

Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Spare space problem.

Post by Morbo »

Hello, guys.
We have two identical 3par 7200 systems in syncrronous replication (FC).
Both have 8 SDD 400GB and 8 SDD 480GB in different shelves. None of those in failed state. Health of both systems is good.
There two CPG on SSD disks (RAID6 (used exclusively for one vvol) and RAID5 (used in ao configuration for all another vvols)) CPGs and vvols on both systems have identical sizes (as all vvols on systems in sync replication))
There is a problem - first system have reserved 30% spare space on each SDD, but second have only 15%. This leads to overusage on SSD disks on first system, these are loaded up to 99% (second system is on 92)
Like this (showpd -space):
One system
Id CagePos Type -State- Size Volume Spare Free Unavail Failed
0 0:0:0 SSD normal 456704 336896 112640 7168 0 0
....
Another system
Id CagePos Type -State- Size Volume Spare Free Unavail Failed
0 0:0:0 SSD normal 456704 322560 62464 71680 0 0


How could we determine why first system reserved twice as much spare size than second system, and how could we recover it?
Thanks.
MammaGutt
Posts: 1578
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Spare space problem.

Post by MammaGutt »

Without knowing anything more than you say here, I would start by looking at «showsys -param» and check SparingAlgorithm parameter
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Re: Spare space problem.

Post by Morbo »

Thank you for reply.
Just checked "showsys -param" on both systems.
Sparing algorithm (as all other params) are exactly the same on both systems.
What additional information could help to determine this issue?

Code: Select all

------Parameter------      --Value---
RawSpaceAlertFC       :          1024
RawSpaceAlertNL       :             0
RawSpaceAlertSSD      :          1024
RemoteSyslog          :             0
RemoteSyslogHost      :       0.0.0.0
SparingAlgorithm      :       Default
EventLogSize          :            3M
VVRetentionTimeMax    :     336 Hours
UpgradeNote           :
PortFailoverEnabled   :           yes
AutoExportAfterReboot :           yes
AllowR5OnNLDrives     :            no
AllowR0               :            no
ThermalShutdown       :           yes
FailoverMatchedSet    :           yes
SessionTimeout        :      01:00:00
HostDIF               :            no
DisableDedup          :            no
PersonaProfile        :    block-only
SingleLunHost         :            no
MammaGutt
Posts: 1578
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Spare space problem.

Post by MammaGutt »

Because I'm lazy and don't have (or remember) the exact chunklet number for 400 and 480 GBs...

Can you share "showpd -c -p -devtype SSD" for both systems?
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Re: Spare space problem.

Post by Morbo »

Sure.
See the difference in spare chunklets.

Code: Select all

3par-1-srvn cli% showpd -c -p -devtype SSD
                             -------- Normal Chunklets -------- ---- Spare Chunklets ----
                             - Used -- -------- Unused -------- - Used - ---- Unused ----
Id CagePos Type State  Total   OK Fail Free Uninit Unavail Fail OK  Fail Free Uninit Fail
 0 0:0:0   SSD  normal   446  324    0   12      0       0    0  0     0  110      0    0
 1 0:1:0   SSD  normal   446  330    0    6      0       0    0  0     0  110      0    0
 2 0:2:0   SSD  normal   446  323    0   14      0       0    0  0     0  109      0    0
 3 0:3:0   SSD  normal   446  328    0    9      0       0    0  0     0  109      0    0
 4 0:4:0   SSD  normal   446  321    0   16      0       0    0  0     0  109      0    0
 5 0:5:0   SSD  normal   446  326    0   12      0       0    0  0     0  108      0    0
 6 0:6:0   SSD  normal   446  322    0   17      0       0    0  0     0  107      0    0
 7 0:7:0   SSD  normal   446  327    0   13      0       0    0  0     0  106      0    0
48 2:0:0   SSD  normal   371  277    0    6      0       0    0  0     0   88      0    0
49 2:1:0   SSD  normal   371  277    0    6      0       0    0  0     0   88      0    0
50 2:2:0   SSD  normal   371  278    0    5      0       0    0  0     0   88      0    0
51 2:3:0   SSD  normal   371  274    0    7      0       0    0  0     0   90      0    0
52 2:4:0   SSD  normal   371  280    0    6      0       0    0  0     0   85      0    0
53 2:5:0   SSD  normal   371  278    0    6      0       0    0  0     0   87      0    0
54 2:6:0   SSD  normal   371  280    0    5      0       0    0  0     0   86      0    0
55 2:7:0   SSD  normal   371  275    0    7      0       0    0  0     0   89      0    0
-----------------------------------------------------------------------------------------
16 total                6536 4820    0  147      0       0    0  0     0 1569      0    0


Code: Select all

3par-2-cpmz cli% showpd -c -p -devtype SSD
                             -------- Normal Chunklets -------- ---- Spare Chunklets ----
                             - Used -- -------- Unused -------- - Used - ---- Unused ----
Id CagePos Type State  Total   OK Fail Free Uninit Unavail Fail OK  Fail Free Uninit Fail
 0 0:0:0   SSD  normal   446  315    0   70      0       0    0  0     0   61      0    0
 1 0:1:0   SSD  normal   446  316    0   69      0       0    0  0     0   61      0    0
 2 0:2:0   SSD  normal   446  317    0   68      0       0    0  0     0   61      0    0
 3 0:3:0   SSD  normal   446  315    0   70      0       0    0  0     0   61      0    0
 4 0:4:0   SSD  normal   446  316    0   69      0       0    0  0     0   61      0    0
 5 0:5:0   SSD  normal   446  316    0   69      0       0    0  0     0   61      0    0
 6 0:6:0   SSD  normal   446  317    0   68      0       0    0  0     0   61      0    0
 7 0:7:0   SSD  normal   446  314    0   71      0       0    0  0     0   61      0    0
48 2:0:0   SSD  normal   371  320    0    0      0       0    0  0     0   51      0    0
49 2:1:0   SSD  normal   371  320    0    0      0       0    0  0     0   51      0    0
50 2:2:0   SSD  normal   371  320    0    0      0       0    0  0     0   51      0    0
51 2:3:0   SSD  normal   371  320    0    0      0       0    0  0     0   51      0    0
52 2:4:0   SSD  normal   371  321    0    0      0       0    0  0     0   50      0    0
53 2:5:0   SSD  normal   371  321    0    0      0       0    0  0     0   50      0    0
54 2:6:0   SSD  normal   371  321    0    0      0       0    0  0     0   50      0    0
55 2:7:0   SSD  normal   371  321    0    0      0       0    0  0     0   50      0    0
-----------------------------------------------------------------------------------------
16 total                6536 5090    0  554      0       0    0  0     0  892      0    0
MammaGutt
Posts: 1578
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Spare space problem.

Post by MammaGutt »

Double check that SparingAlgorithm is set to default on 3par-1-srvn.

If that is set to default, run "admithw" on the system to recreate spare capacity. The correct number of spare chunklets should be 892 (similar to the capacity of 2 drives per 24 SSDs).

Are both running the same 3PAR OS version?
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Re: Spare space problem.

Post by Morbo »

Yes, systems are identical.
The only difference is that the system with normal spare size have adaptive flash cache enabled.
It went like this:
We`d decided to try to use AFC on both systems.
We succefully enabled it (256GB in size) on one system.
But on the other, it failed to enable with error "Could not find enough available disk space."
After that we started to investigate and found overusage of spare space on this system.

I`ve run admithw on 3par, but it has no effect.
Should i run tunesys also?
MammaGutt
Posts: 1578
Joined: Mon Sep 21, 2015 2:11 pm
Location: Europe

Re: Spare space problem.

Post by MammaGutt »

Morbo wrote:Yes, systems are identical.
The only difference is that the system with normal spare size have adaptive flash cache enabled.
It went like this:
We`d decided to try to use AFC on both systems.
We succefully enabled it (256GB in size) on one system.
But on the other, it failed to enable with error "Could not find enough available disk space."
After that we started to investigate and found overusage of spare space on this system.

I`ve run admithw on 3par, but it has no effect.
Should i run tunesys also?


If you compare "showpd -c" on both systems. Is it only the SSDs that have double spare capacity or is it the same for all drive types?

You could try and do "setsys SparingAlgorithm Minimal" and admithw again. It should reduce (but it should reduce it more than you want). Reduced sparing will allow for less hardware failure before you risk data loss.
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.
Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Re: Spare space problem.

Post by Morbo »

Code: Select all

   Type   Drives   Total chunks   Spare(3par-1)   Spare(3par-2)
FC         32   30960               1340         2232
NL         24   86664               4297         4990
SSD         16    6536               1569          892


On second system there is similar situation but with FC drives.
Morbo
Posts: 11
Joined: Thu Apr 20, 2023 1:24 am

Re: Spare space problem.

Post by Morbo »

If I do like this:

Code: Select all

setsys SparingAlgorithm Minimal
admithw

setsys SparingAlgorithm Default
admithw


Will it help restore default spare chunks count?
Post Reply