1

Topic: The strange self-liquidation of basis

All kind time of days!
Happens terrible - in the middle of the working day the basis 1 fell.
Fell strange - it was cleared completely. Does not remain any user object. As though from basis deleted all of them. Thus removal actually bases was not - she lived on the same files.mdf and.ldf
All right, have been normally adjusted , tore by the test machine, discovered about the moment of falling and basis recovered in a state for pair minutes before falling. Recovered simply rollback by the necessary moment of time.
But all would be desirable to establish the reason of such falling. To find that request which managed to fill up basis to which there were many the active connections. Hands did not possible to turn such feint.
There is full , there is a copy of a broad gull of transactions. The crash moment is known to within pair minutes.
Question. What means/methods it is possible to establish actually the reason of this crash?
In an ideal  (as in a debugger) to transit all requests which on basis were executed before falling
---------------------
I always lie.

2

Re: The strange self-liquidation of basis

Contrast;
In   precisely is not present DROP TABLE?
With you remained broken.mdf and.ldf? What speaks DBCC CHECKDB?

3

Re: The strange self-liquidation of basis

Basically Mind already wrote that you should look.
Whether it is included Default Trace - here here short that it
If there was a broad gull (if you on the full model certainly), look fn_dblog if this moment already got in Backup, fn_dump_log. An example here

4

Re: The strange self-liquidation of basis

Contrast wrote:

All kind time of days!
Happens terrible - in the middle of the working day the basis 1 fell.
Fell strange - it was cleared completely. Does not remain any user object.
as though from basis deleted all of them. Thus removal actually bases was not - she lived on the same files.mdf and.ldf
All right, have been normally adjusted , tore by the test machine, discovered about the moment of falling and basis recovered in a state for pair minutes before falling. Recovered simply rollback by the necessary moment of time.
But all would be desirable to establish the reason of such falling. To find that request which managed to fill up basis to which there were many the active connections. Hands did not possible to turn such feint.
There is full , there is a copy of a broad gull of transactions. The crash moment is known to within pair minutes.
Question. What means/methods it is possible to establish actually the reason of this crash?
In an ideal  (as in a debugger) to transit all requests which on basis were executed before falling
---------------------
I always lie .

.

5

Re: The strange self-liquidation of basis

Contrast;
so to delete all objects the hand is tired unless to kill all , and then ... So yes, DEFAULT TRACE to you all tells

6

Re: The strange self-liquidation of basis

TaPaK;
And all of them automated

7

Re: The strange self-liquidation of basis

Someone banged basis, and then created the empty

8

Re: The strange self-liquidation of basis

wrote:

someone banged basis, and then created empty

Writes:

wrote:

Thus removals actually bases were not - she lived on the same files.mdf and.ldf

I personally understand "on the same" is when Date created centenary prescription.
But it is possible certainly and to specify once again: what there at us with date of creation of files?
Also what in dbi_crdate y dbcc dbinfo with tableresults?

9

Re: The strange self-liquidation of basis

Yasha123 wrote:

I personally understand "on the same" is when Date created centenary prescription.
But it is possible certainly and to specify once again: what there at us with date of creation of files?
Also what in dbi_crdate y dbcc dbinfo with tableresults?

dbi_crdate Gives 1900 year. But it, probably is not important - files old. Creation date corresponds to the expected.
With Default Trace was late - it was re-recorded...
I will smoke fn_dump_dblog...  here at once did not give in...
Thanks all huge!
The ray of hope is. You look, I will dig out

10

Re: The strange self-liquidation of basis

Contrast wrote:

With Default Trace was late - it was re-recorded...

And precisely searched on all to five files?
after all on one leaking

11

Re: The strange self-liquidation of basis

Yasha123 wrote:

and precisely searched on all to five files?
after all on one leaking

If it is fair, I do not know on what searched. But all five files have creation date fresh, than crash date.