Org mode as an helper tool for my job position (DBA)

Org mode as an helper tool for my job position (DBA)

[Xavier Maillard]

Hi,

I am a DBA (database administrator) with really tons of database (more than 200)
to "survey".

As such, I have written many tools for various tasks. Currently these tools
produce logs (tons of) in many different formats. That's not very homogeneous
but that's something I want to remedy.

My current (en)vision is quite simple: outlined document. That's where org-mode
enters the dance :)

Currently reading/parsing/getting the right information from each tools is
rather time consuming. Having org-mode based log format would speed this by a
factor of 1000 (at least).

Let take the example of an update statistics. What informations are really
relevant for me is simply this:

1. server name
2. start date/end date
3. return code
4. database name

Then for each table/index of the given database, I need this:

1. object name
2. status
3. time spent

Each object would have extra informations as a note under the entry.

Producing the right org-mode file structure from any script/c code is rather
simple, in fact, I have done this as a quick and dirty hack for one tool.

Where I am failing is in having a column view of the document (I want to be able
to only view time spent for each object name to measure the performance of the
last update statistics run).

What would I need to add for each of the table/index "todo" entry to have this
information at hand ?

As an example, updating index statistics will present an org file like this one:

#title Update statistics of sql server

* General information
- start date
- end date
- return code

* Index statistics (I want column view from here for each object name)
** Index name 1
informations for index name 1 go here (time spent, etc...)
** Index name 2
* Index name <n>
...

I hope I have been clear :)

Thank you for your feedback.

Regards

Xavier

Re: Org mode as an helper tool for my job position (DBA)

[Bernt Hansen]

Xavier Maillard <xma@gnu.org> writes:

> Hi,
>

Hi!

> Producing the right org-mode file structure from any script/c code is rather
> simple, in fact, I have done this as a quick and dirty hack for one tool.
>
> Where I am failing is in having a column view of the document (I want to be able
> to only view time spent for each object name to measure the performance of the
> last update statistics run).
>
> What would I need to add for each of the table/index "todo" entry to have this
> information at hand ?
>
> As an example, updating index statistics will present an org file like this one:
>
> #title Update statistics of sql server
>
> * General information
> - start date
> - end date
> - return code
>
> * Index statistics (I want column view from here for each object name)
> ** Index name 1
> informations for index name 1 go here (time spent, etc...)
> ** Index name 2
> * Index name <n>
> ...
>

Here's a sample org file.  Just add properties to the tasks (objects)
you need in the column view and a #+COLUMNS: line describing what fields
to include in the columns.

,----[ Sample org file ]
| #+STARTUP:
| #+COLUMNS: %40ITEM(Task) %30Status(Status) %30TimeSpent(Time Spent){:}
| * General information
| - start date
| - end date
| - return code
| 
| * Index statistics (I want column view from here for each object name)
| ** Index name 1
| informations for index name 1 go here (time spent, etc...)
| :PROPERTIES:
| :TimeSpent: 00:30
| :Status: S42
| :END:
| ** Index name 2
| :PROPERTIES:
| :TimeSpent: 00:17
| :Status: R19
| :END:
| * Index name <n>
| :PROPERTIES:
| :TimeSpent: 02:16
| :Status: 42
| :END:
`----

Hope that helps,

Bernt

Re: Org mode as an helper tool for my job position (DBA)

[Russell Adams]

> Let take the example of an update statistics. What informations are really
> relevant for me is simply this:
> 
> 1. server name
> 2. start date/end date
> 3. return code
> 4. database name
> 
> Then for each table/index of the given database, I need this:
> 
> 1. object name
> 2. status
> 3. time spent
> 
> Each object would have extra informations as a note under the entry.
> 
> Producing the right org-mode file structure from any script/c code is rather
> simple, in fact, I have done this as a quick and dirty hack for one tool.
> 
> Where I am failing is in having a column view of the document (I want to be able
> to only view time spent for each object name to measure the performance of the
> last update statistics run).

It sounds like you should explore the property drawer and properties,
using Org's column view. It can summarize properties across items,
which is very handy whether turning out a budget, timeline, etc.

I think you're right, that Org could make reviewing this kind of data
very fast.

Here's a short example you can use with column view (C-c C-x C-c):

*  Jobs
   :PROPERTIES:
   :COLUMNS: %35ITEM %8TODO %5DURATION %2EXITCODE
   :END:

**  Job 1
   :PROPERTIES:
   :STARTTIME: 1:00
   :ENDTIME:  2:00
   :EXITCODE: 0
   :DURATION: 1:00
   :END:

Log from job goes here.

**  Job 2
   :PROPERTIES:
   :STARTTIME: 2:00
   :ENDTIME:  3:30
   :EXITCODE: -1
   :DURATION: 1:30
   :END:

Log from job goes here.


------------------------------------------------------------------
Russell Adams                            RLAdams@AdamsInfoServ.com

PGP Key ID:     0x1160DCB3           http://www.adamsinfoserv.com/

Fingerprint:    1723 D8CA 4280 1EC9 557F  66E8 1154 E018 1160 DCB3