[Dspam-user] MySQL 5.7 is evil

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

[Dspam-user] MySQL 5.7 is evil

Marco
Hello,

  days ago I switched from MySQL5.6 to MySQL5.7, fresh DB, InnoDB,  
dspam with TOE.

DSPAM stops to work for a tons of "lock wait timeout". Back to MySQL  
5.6 and DSPAM returns to work greatly. No differences on hardware and  
resources. Only the DBMS engine changed.

I think that MySQL5.7 introduces something that DSPAM doesn't like.  
Does anyone experience this problem?

Thank you
Best Regards
Marco


------------------------------------------------------------------------------
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user
Reply | Threaded
Open this post in threaded view
|

Re: [Dspam-user] MySQL 5.7 is evil

Chris Cappuccio
Marco [[hidden email]] wrote:

> Hello,
>
>   days ago I switched from MySQL5.6 to MySQL5.7, fresh DB, InnoDB,  
> dspam with TOE.
>
> DSPAM stops to work for a tons of "lock wait timeout". Back to MySQL  
> 5.6 and DSPAM returns to work greatly. No differences on hardware and  
> resources. Only the DBMS engine changed.
>
> I think that MySQL5.7 introduces something that DSPAM doesn't like.  
> Does anyone experience this problem?
>

I think you should look at upgrading to MariaDB and not Oracle MySQL.



------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user
Reply | Threaded
Open this post in threaded view
|

Re: [Dspam-user] MySQL 5.7 is evil

P.V.Anthony-2
On 04/20/2016 12:40 AM, Chris Cappuccio wrote:

> I think you should look at upgrading to MariaDB and not Oracle MySQL.

The other option is to use postgresql.

P.V.Anthony



------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user

smime.p7s (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [Dspam-user] MySQL 5.7 is evil

Phil Stracchino
In reply to this post by Marco
On 04/06/16 06:07, Marco wrote:

> Hello,
>
>   days ago I switched from MySQL5.6 to MySQL5.7, fresh DB, InnoDB,  
> dspam with TOE.
>
> DSPAM stops to work for a tons of "lock wait timeout". Back to MySQL  
> 5.6 and DSPAM returns to work greatly. No differences on hardware and  
> resources. Only the DBMS engine changed.
>
> I think that MySQL5.7 introduces something that DSPAM doesn't like.  
> Does anyone experience this problem?


Honestly, if you're seeing lock wait timeouts the first thing to do is
check for a tuning problem.  This includes making sure you aren't using
MyISAM tables.


--
  Phil Stracchino
  Babylon Communications
  [hidden email]
  [hidden email]
  Landline: 603.293.8485

------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user
Reply | Threaded
Open this post in threaded view
|

[Dspam-user] Using hash storage driver - cssclean fails

Paul Goyette
In reply to this post by Chris Cappuccio
I've got a real small set-up (two users, total of maybe 500 messages per
day), and really have no justification for adding complexity (and the
learning curve) for sqlite or similar.  So I'm just using the simple
hash storage driver.

Everything seems to run fine, except for the database maintenance.  I
have a couple of issues with cssclean and csscompress:

* Both utilities seem to require an absolute path specification of the
   file that is being maintained.  Otherwise, when the utility is
   finished and it tries to rename the work file, it fails with EXDEV
   (attempt to create hard link to a file on another file system).

* cssclean starts to run, takes several seconds, and then failes with
   "Floating exception".  I did a little bit of exploration in the
   source code, and discovered that this really isn't a FPE error, but
   rather it is the result of an integer divide-by-zero.  (This is on
   a NetBSD system...)

The first issue I can manage, but I have not been able to understand the
source well enough to make any progress on the second issue.

Has anyone else seen this?  And maybe have a solution?  Indeed, is
anyone else even using the hash storage driver?

Thanks in advance for any suggestions or guidance.   :)



+------------------+--------------------------+------------------------+
| Paul Goyette     | PGP Key fingerprint:     | E-mail addresses:      |
| (Retired)        | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com   |
| Kernel Developer | 0786 F758 55DE 53BA 7731 | pgoyette at netbsd.org |
+------------------+--------------------------+------------------------+

------------------------------------------------------------------------------
Find and fix application performance issues faster with Applications Manager
Applications Manager provides deep performance insights into multiple tiers of
your business applications. It resolves application problems quickly and
reduces your MTTR. Get your free trial!
https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user
RW
Reply | Threaded
Open this post in threaded view
|

Re: [Dspam-user] Using hash storage driver - cssclean fails

RW
On Wed, 20 Apr 2016 06:42:07 +0800 (PHT)
Paul Goyette wrote:

> I've got a real small set-up (two users, total of maybe 500 messages
> per day), and really have no justification for adding complexity (and
> the learning curve) for sqlite or similar.  So I'm just using the
> simple hash storage driver.

IIRC there is no learning curve for just switching to sqlite, it's just
a matter of editing the StorageDriver setting, and then rebuilding from
corpus or starting over.

There may be a bit of a learning curve to purging on sqlite, but
cssclean hasn't worked for so long that it probably never will.


> Everything seems to run fine, except for the database maintenance.  I
> have a couple of issues with cssclean and csscompress:
>
...
> Has anyone else seen this?  And maybe have a solution?  Indeed, is
> anyone else even using the hash storage driver?

When I tried sqlite I immediately switched back. I like to try
different settings and that often benefits from, or even requires,
retraining from corpus. This take 2-3 minutes with the hash driver, but
over 7 hours with sqlite.

I just add anything I train with into the corpus I used for initial
training. Occasionally I trim the corpus and recreate the database.

With low-volume TOE you could probably just ignore it. My understanding
is that it's  mostly csscompress that keep the file fast, and that
works.


------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Dspam-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/dspam-user