Adding used disks from a retiered T400 to another T400

[warlock]

Hi Guys,

First time here so hopefully someone can help me out,

We have 2 T400 both out of maintenance, and 1 has been decommissioned, we are trying use the disk on the decomm'd Array in the Running array,

We preformed
servicemag start 4 0

once it succeeded
we took the disk out of the unused array and installed in the place of the failed disk, then
servicemag start 4 0

and this completed, but only the originall good drives came online and the failed disk reports the following

Code: Select all

T400INSERV01 cli% showpd -c 184
                              ------- Normal Chunklets -------- ---- Spare Chunklets ----
                              - Used - -------- Unused -------- - Used - ---- Unused ----
 Id CagePos Type State  Total OK  Fail Free Uninit Unavail Fail OK  Fail Free Uninit Fail
184 4:0:2?  FC   failed  1115  0     0    0      0     571  544  0     0    0      0    0
-----------------------------------------------------------------------------------------
  1 total                1115  0     0    0      0     571  544  0     0    0      0    0
T400INSERV01 cli% showpd -state 184
 Id CagePos Type -State- ----------------------------------Detailed_State----------------------------------
184 4:0:2?  FC   failed  vacated,missing,not_available_for_allocations,invalid_media,failed_hardware,spinup
-----------------------------------------------------------------------------------------------------------
  1 total

How can I force the disk to come on-line and be accepted by the array, remembering that this disk was once used in another array and was never dismissed from that array.

Thanks.

[RitonLaBevue]

Having an high end array without support is just... s t u p i d.

[warlock]

Having an high end array without support is just... s t u p i d.

I'll take this comment with a pinch of salt, and educate you a bit,

The unit is going on for 5 years old, we are using it for non-critical dev & test, it was replaced with an all flash array, but rather than through a good piece and equipment worth over $1mil away we re-purposed it for POC, DEV&TEST, and RND. Now HP will not work on the unit unless it has an active support contract and as the unit is perfect except for a few failed disks, why not maintain it ourselves and ask on this forum if anyone can assist, after all if everyone had support why would we even bother with forums likes this ?

[RitonLaBevue]

I now how old are T series. I started in august 2008 on 2 T400.

On these old arrays disks replacement is done by support not by end customers as it is done on 7xx0.
Thanks for the education

I'm not in charge of support in my company but the few failed disks i paid attention to never had so many errors...

One thing you could do: insert old mag in your active array, do admithw, check if the disk are in a good state, remove the spare from them (removespare -f pdid:a) and dismiss the pd (dismisspd pdid). That way you will be sure the disks you use are valids.

For the trouble you are facing i got no idea.
You can ever try this :
showpd -i
controlport chmederr set -1 pdwwn
controlport chmederr set -2 pdwwn
controlport chmederr unset -2 pdwwn
controlport chmederr unset -1 pdwwn
controlport clearerr pdwwn

That will force the reset of some errors...

But... I did not know forums was made to replace failed disks.

[warlock]

Ok so it seems the replacement disks were actually accepted, it's just the old disks were not removed and are still reproting in the system

Code: Select all

T400INSERV01 cli% showpd -p -mg 0 -cg 4
                            ---Size(MB)--- ----Ports----
 Id CagePos Type RPM State    Total   Free A      B      Cap(GB)
112 4:0:0   FC    15 normal  285440 114688 2:2:1* 3:2:1      300
113 4:0:1?  FC    15 failed  285440      0 -----  -----      300
115 4:0:3   FC    15 normal  285440 117760 2:2:1  3:2:1*     300
149 4:0:1   FC    15 normal  285440 143872 2:2:1  3:2:1*     300
151 4:0:2   FC    15 normal  285440 160256 2:2:1* 3:2:1      300
184 4:0:2?  FC    15 failed  285440      0 -----  -----      300
----------------------------------------------------------------
  6 total                   1712640 536576

I assume that I would just need to run ?

Code: Select all

dismisspd 113
dismisspd 184

[RitonLaBevue]

running these commands is harmfull...
If the pds still have data (spare, etc...) you ll know it.

[warlock]

Turns out running those commands was the right thing to do,

From this link
http://ftp1.osiatis.com:8081/doc.do;jsessionid=1og68mt5p5onj?docId=p_dvd12%2Femr_na-c02660861.pdf

Code: Select all

For the servicemag start command only, instead of using the <cage_ID> <magazine>
specifiers, the disk(s) to be serviced can be specified using the -pdid option and servicemag 
resume will automatically dismiss the specified disk(s) when it completes. However, if 
servicemag unmark <cage> <mag> is run before the servicemag resume command, the 
disk(s) specified with the –pdid option, will not be automatically dismissed.
X Workaround - When the servicemag resume <cage> <mag> has successfully completed, 
check using the showpd command if the disks specified with the -pdid option are still in the 
system. If so, you will need to dismiss those disks manually by running dismisspd <pdid>.

We didn't use the pdid instead we did the cage and mag, and this is way the old disks remained registered even though they had been successfully replaced.