dbqp.py needs better naming for shared memory symlinks

Bug #708865 reported by Patrick Crews
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Drizzle
Fix Released
Low
Patrick Crews
7.0
Fix Released
Low
Patrick Crews

Bug Description

Need better naming / management of shared memory working directories we create.

Related branches

Changed in drizzle:
status: New → Confirmed
importance: Undecided → Low
assignee: nobody → Patrick Crews (patrick-crews)
milestone: none → 2011-01-31
Revision history for this message
Patrick Crews (patrick-crews) wrote :

Preparing branch that will create directories named dbqp_workdir_$username.
This will help prevent issues with permissions / user-user collisions.
Will need to investigate build-thread or other options later.

Changed in drizzle:
status: Confirmed → In Progress
Revision history for this message
Stewart Smith (stewart) wrote : Re: [Bug 708865] Re: dbqp.py needs better naming for shared memory symlinks

On Fri, 28 Jan 2011 20:12:07 -0000, Patrick Crews <email address hidden> wrote:
> Preparing branch that will create directories named dbqp_workdir_$username.
> This will help prevent issues with permissions / user-user collisions.
> Will need to investigate build-thread or other options later.

Why not completely skip it and just use libeatmydata? We can even
preload it on OSX (if we get the foo correct, but it is possible). I
have not seen any situation where in any practical sense a tmpfs is
faster than libeatmydata (at least on Linux, and who cares about
anything else... besides, tmpfs is harder to find on others :).

--
Stewart Smith

Changed in drizzle:
status: In Progress → Fix Committed
Revision history for this message
Patrick Crews (patrick-crews) wrote :

libeatmydata is definitely on my radar. However, there are currently
several moving parts that we're still polishing / testing and using tmpfs
was simply quicker to have up and running : )

Since the code has been refactored, it won't be difficult to add
libeatmydata later on, but at the moment, we are going for what works with
minimal effort and thought to get the big picture better set.

On Fri, Jan 28, 2011 at 7:53 PM, Stewart Smith <email address hidden>wrote:

> On Fri, 28 Jan 2011 20:12:07 -0000, Patrick Crews <
> <email address hidden>> wrote:
> > Preparing branch that will create directories named
> dbqp_workdir_$username.
> > This will help prevent issues with permissions / user-user collisions.
> > Will need to investigate build-thread or other options later.
>
> Why not completely skip it and just use libeatmydata? We can even
> preload it on OSX (if we get the foo correct, but it is possible). I
> have not seen any situation where in any practical sense a tmpfs is
> faster than libeatmydata (at least on Linux, and who cares about
> anything else... besides, tmpfs is harder to find on others :).
>
> --
> Stewart Smith
>
> --
> You received this bug notification because you are a direct subscriber
> of the bug.
> https://bugs.launchpad.net/bugs/708865
>
> Title:
> dbqp.py needs better naming for shared memory symlinks
>
> Status in A Lightweight SQL Database for Cloud Infrastructure and Web
> Applications:
> Fix Released
> Status in Drizzle elliott series:
> Fix Released
>
> Bug description:
> Need better naming / management of shared memory working directories
> we create.
>
> To unsubscribe from this bug, go to:
> https://bugs.launchpad.net/drizzle/+bug/708865/+subscribe
>

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.