1

Topic: Remained lifetime SSD

Impart experience: how you on  supervise remained lifetime SSD?
My problem:
SQL 2008, 64, Ent, translated bases and system on SSD, now we think when they (SSD) die?
Therefore it would be desirable to consider TBW with SSD disks.
Singularity: disks in an array,  from from an array do not read the majority.
The controler: Intel Integrated Raid Controller RMS3CC80, BBU
- Whether there is a possibility to subtract directly from SQL?
- Whether is , it is possible paid.

2

Re: Remained lifetime SSD

Alexander Us;
At you all problems in basis optimization are solved, and now it is possible a hogwash ?))

3

Re: Remained lifetime SSD

StarikNavy;
As SSD in an array have identical resource TBW fall out simultaneously.
At people the such happened much earlier, than they could present to themselves.
I consider such check necessary.

4

Re: Remained lifetime SSD

Some companions too die much earlier, than they could present to themselves.
As though to learn exact date of the future inevitable death, it is desirable using SQL?

5

Re: Remained lifetime SSD

StarikNavy wrote:

Alexander Us;
At you all problems in basis optimization are solved, and now it is possible a hogwash ?))

Any actions with a software do not cancel iron administration. UPS, cooling, a state of disks need to be supervised even at nonoptimal bases.

Alexander Us wrote:

Therefore it would be desirable to consider TBW with SSD disks.

And unless in SMART there is no this information?
it is necessary for you  administration of this controler.

6

Re: Remained lifetime SSD

Alexander Us;
Once a month substitute one disk from an array on new,
connect to a computer and native utilities look as it badly. Accordingly already there is an information on the remained.
Then change backwards, or do not change. :-)

7

Re: Remained lifetime SSD

vborets wrote:

Alexander Us;
Once a month substitute one disk from an array on new,
connect to a computer and native utilities look as it badly. Accordingly already there is an information on the remained.
Then change backwards, or do not change. :-)

Yes in dock to the controler it is written that the smart of disks is accessible
It is not necessary such complexities...

8

Re: Remained lifetime SSD

alexeyvg wrote:

Yes in dock to the controler it is written that the smart of disks is accessible
It is not necessary such complexities...

alexeyvg;
Thanks for support though one person on a branch understands that for check of level of oil can be the indicator.
And then some consider that  it is not necessary to check oil;
Others advise  the motor and to look eyes.
There was a question:
Who advises tested  which works with this controler: (Intel Integrated Raid Controller RMS3CC80, BBU)
Or it is possible as easily that to read out without ?

9

Re: Remained lifetime SSD

Alexander Us wrote:

Or it is possible as easily that to read out without ?

Generally all salt of that council not in "read out", and in "periodically change". That is in your array there should not be disks with equally worked out resource, always there should be the oldest which fails the first, but makes it alone. And it is better - RAID 6 to endure failure at once two.

10

Re: Remained lifetime SSD

Alexander Us wrote:

though one person on a branch understands that for check of level of oil there can be an indicator.

Not absolutely clearly, at what here the server?
After all and date of death  it is desirable to know beforehand, differently the server suffers.
About iron probably it is necessary to ask in iron section;
No less than about analyses/health to be interested in more suitable place

11

Re: Remained lifetime SSD

Dimitry Sibiryakov;
Your council "periodically "  when is a question of 1-2 disks.
If it is more than them, it is necessary to get the gnome who will be engaged in changeover of disks.
Not to mention ineffective expending of means for disks for changeover.
In an ideal it would be desirable to have the notification that disk XYZ settled the TBW a resource on, for example, 80 %.
Then and to change.

12

Re: Remained lifetime SSD

Alexander Us;
Regular Intel SSD Datacenter Tool, for example
https://downloadcenter.intel.com/download/27144?v=t
But it supports only Intel SSD and only Enterprise series.

13

Re: Remained lifetime SSD

If  not to do, yes, it is possible to sit, shiver.

14

Re: Remained lifetime SSD

a_shats wrote:

Alexander Us;
Regular Intel SSD Datacenter Tool, for example
https://downloadcenter.intel.com/download/27144?v=t
But it supports only Intel SSD and only Enterprise series.

Thanks, alas disks though Enterprise but not from Intel and from Samsung.

15

Re: Remained lifetime SSD

Kolosov wrote:

If  not to do, yes, it is possible to sit, shiver.

And at what here ???
Even at their presence if at you ALL array * takes off that you will do?
*Well or separate disks, but one after another so quickly that you stupidly have not time to change them.

16

Re: Remained lifetime SSD

Alexander Us wrote:

you will do that?

I Will switch to the second .

17

Re: Remained lifetime SSD

Sergey Alekseevich wrote:

it is passed...
I will switch to the second .

That means most likely short-term but failure in operation of applications.
Supervising TBW it it would be possible to avoid.

18

Re: Remained lifetime SSD

Sergey Alekseevich wrote:

it is passed...
I will switch to the second .

Excuse that I will get, but to me it is dispute it is interesting, a little from the hypothetical point of view.
And  that you plan to switch to the second? Reserve ?

19

Re: Remained lifetime SSD

At Gavrilenko not hypothetical, but most that on is real Always On.
I.e. all  on generally other server with an own array

20

Re: Remained lifetime SSD

o-o wrote:

at Gavrilenko not hypothetical, but most that on is real Always On.
I.e. all  on generally other server with an own array

And switching on the second  happens for application imperceptibly, without noticeable  or breakaways ?

21

Re: Remained lifetime SSD

To realize that TBW this guessing on a crystal full-sphere (moreover and  as hardly you find the normal utility  in an array), and here hot spare in raid', moreover and with a disk at which certainly smaller TBW that is necessary. Without harm to a production environment of times in a month \quarter \half a year \year the disk hot spare is excepted and  on a normal computer by the native utility, interposed reversely. All sleep easily. Disks in raid will work on 100 % of a resource, do not die a natural death yet, instead of will go to breakage ahead of time.

22

Re: Remained lifetime SSD

the Gander wrote:

to Realize that TBW this guessing on a crystal full-sphere (moreover and  as hardly you find the normal utility  in an array), and here hot spare in raid', moreover and with a disk at which certainly smaller TBW that is necessary

Perfectly, too I do not understand accent on TBW.
Declared by vendor TBW generally in tens times more, than real, and it (real) unequal for different copies.
So all disks do not take off at once. And jotas-svap the disks specially for this purpose invented, quite help.
That does not cancel desirability of monitoring because to understand resource exhaustion it is possible on other parameters.
Yes, also it is necessary to be ready immediately to buy disks at switching of jotas-svapa, not to expand it for the period of semi-annual coordination of the budget.
In general, total: jotas-svap, readiness to change disks, monitoring of notifications from an array, well and  - to search for utilities.

23

Re: Remained lifetime SSD

alexeyvg wrote:

Declared by vendor TBW generally in tens times more, than real

in tens times it is less, of course, instead of more.

24

Re: Remained lifetime SSD

Alexander Us wrote:

it is passed...
And switching on the second  happens for application imperceptibly, without noticeable  or breakaways ?

Esteem about AlwaysOn. It can help not only when suddenly all disks in an array took off, but at a heap of other failures which are not repaired by banal changeover of a disk (for example, because of precipitates the cooling system in  broke or mother on the server burned down).

25

Re: Remained lifetime SSD

o-o wrote:

at Gavrilenko not hypothetical, but most that on is real Always On.
I.e. all  on generally other server with an own array

that AOn at companion Gavrilenko I am real I do not doubt smile
To me what exactly was simple is interesting was implied by switching.
And why I so asked - in my case on storage system that can hang much and not all the data, storable on  are controlled by a sequel, and from  bases, bases use not all AOn.
And analysis of an array because of embarkation of several SSD would be an unpleasant adventure as protection against such affairs I see only in  all array or its separate pools means  in a reserve data-center. And this business rather  - allows it can not only only all. And the indicator of state SSD about which asks the HARDWARE - a piece, in my opinion, useful.
Sergey Alekseevicha's answer is clear, thanks