Hey
I took a quick look at the patch and noticed you removed the dedup at the
begining.
Are you using postgresql or oracle?
I know from testing that on a properly tuned postgresql server that initial
query reduces a large number of querys latter on resyncs each of which take
almost as much time as that initial query.
Now on the initial sync it does cause it to take a little longer but the
difference should be negligable.
The only exception to this is if postgresql isn't given enough working
memory then it would be a significant choke point but then many other
common queries against the database would be slow as well.
On Oct 24, 2012 6:14 PM, "Paul Robert Marino" <[email protected]> wrote:

> I'm on vacation this week but ill take a look next week. I can tell you
> I'm sure taking out the spacewalk source cappability would speed it up but
> I have reasons I wanted that capability so I have no intentino of removing
> it but I could streamline it a little with an additional flag and 2 to 3
> more subroutines
> On Oct 23, 2012 8:00 AM, "Franky Van Liedekerke" <[email protected]>
> wrote:
>
>> **
>>
>> On 2012-10-15 15:17, Franky Van Liedekerke wrote:
>>
>> Yet another errata clone script ... for CentOS and RedHat
>>
>>
>> For CentOS:
>> - first some shell calls to get the latest announces from the centos
>> archive
>>   (but not by scraping the announces list, but getting digests, much less
>>   traffic then)
>>   You can change the number of announces anyway you want by changing the
>>   wget command to your liking
>> - then the perl script comes along, parsing the digest files and looking
>> in 1
>>   channel (yes, one) for package availability and creating the errata
>> there.
>>   It has optional integration with RHN for notes, description, topic
>> info, and
>>   CVE's and/or can use the OVAL file like the errata-import.pl script
>> does.
>>   The created errata gets a suffix based on the OS version and
>> architecture
>>   (e.g. ":C5-64" or "C6-32"), because the same errata can exist for
>> multiple
>>   OS versions and architectures (and creating the errata for more than one
>>   base channel would result in packages being copied which is a mess
>> again).
>>   Also a proxy can be defined for spacewalk and or RHN servers
>> Or, for RedHat:
>> - log in to RHN, get the errata for the specified channel
>>   (possibility to define the date range)
>> - then follow the same logic as for CentOS
>>
>>
>>
>> For the persons interested in using my script:
>>
>> - I confirmed it to be working for RHEL4 (Extended support), RHEL5, RHEL6
>> errata and for CentOS 5 and 6 (32 and 64bit)
>> - something I forgot to mention in the original mail: proxy support is in
>> there (just not authenticated, but that can be added if wanted)
>>
>> See github for the latest version (some small bugfixes have been made):
>> https://github.com/liedekef/spacewalk_scripts
>>
>> Feedback is greatly appreciated!
>>
>> Franky
>>
>>
>> _______________________________________________
>> Spacewalk-list mailing list
>> [email protected]
>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>
>
_______________________________________________
Spacewalk-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/spacewalk-list

Reply via email to