On Thu, Mar 5, 2009 at 4:06 PM, Nick Stolwijk <[email protected]> wrote:
> The query originally was a freemarker template, something like
> <d:literal>${querytext}*</d:literal> where the querytext was empty, a
> bug in our application. We just found it weird behaviour it could
> bring down the whole repository.

Yes, i agree the repository should not be brought down!  A pid file
would be helpful if you want the issue investigated,

Regards

>
> With regards,
>
> Nick Stolwijk
> ~Java Developer~
>
> Iprofs BV.
> Claus Sluterweg 125
> 2012 WS Haarlem
> www.iprofs.nl
>
>
>
> On Thu, Mar 5, 2009 at 4:02 PM, Ard Schrijvers
> <[email protected]> wrote:
>> On Thu, Mar 5, 2009 at 3:55 PM, Nick Stolwijk <[email protected]> 
>> wrote:
>>> We've encountered a problem with our dasl query, such that we can let
>>> the repository crash. It seems it goes wrong when executing a query
>>> which contains a <d:contains>*</d:contains>.
>>>
>>> I am running against Hippo Repository 1.2.15.1. I have written a
>>> simple testprogram using the Hippo Repository Adapter 2.12.00-dev (due
>>> to a concurrency bug in 2.11). My testprogram creates 20 threads which
>>> are going to execute the next query over and over having a little
>>> sleep between each execution.
>>>
>>> The query:
>>>
>>> <?xml version="1.0" encoding="UTF-8"?>
>>> <d:searchrequest xmlns:s="http://jakarta.apache.org/slide/";
>>> xmlns:h="http://hippo.nl/cms/1.0"; xmlns:d="DAV:">
>>>    <d:basicsearch>
>>>        <d:select>
>>>            <d:prop>
>>>                <s:nrHits />
>>>                <h:caption />
>>>                <h:index />
>>>            </d:prop>
>>>        </d:select>
>>>        <d:from>
>>>            <d:scope>
>>>                <d:href>content/</d:href>
>>>                <d:depth>Infinity</d:depth>
>>>            </d:scope>
>>>        </d:from>
>>>        <d:where>
>>>            <d:contains>*</d:contains>
>>>        </d:where>
>>>        <limit xmlns="DAV:">
>>>            <nresults>12</nresults>
>>>            <offset xmlns="http://jakarta.apache.org/slide/";>0</offset>
>>>        </limit>
>>>        <d:orderby>
>>>            <d:order>
>>>                <d:prop>
>>>                    <h:index />
>>>                </d:prop>
>>>                <d:ascending />
>>>            </d:order>
>>>        </d:orderby>
>>>    </d:basicsearch>
>>> </d:searchrequest>
>>>
>>> It can crash the repository running with 512 MB memory, but with 1 GB
>>> it is also possible. (It takes just a minute longer or so)
>>
>> Well, the query ofcourse is strange to do. Most likely the repository
>> crashes because of lucene is loading all terms in memory. It thus also
>> depends on the size of your repository. If you take a look at a pid
>> file i suspect you seeing all memory beeing eating up by lucene.
>>
>> What are you trying to achieve anyway?
>>
>> Ard
>>
>>>
>>> Is this expected behaviour?
>>>
>>> Nick Stolwijk
>>> ~Java Developer~
>>>
>>> Iprofs BV.
>>> Claus Sluterweg 125
>>> 2012 WS Haarlem
>>> www.iprofs.nl
>>> ********************************************
>>> Hippocms-dev: Hippo CMS development public mailinglist
>>>
>>> Searchable archives can be found at:
>>> MarkMail: http://hippocms-dev.markmail.org
>>> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html
>>>
>>>
>> ********************************************
>> Hippocms-dev: Hippo CMS development public mailinglist
>>
>> Searchable archives can be found at:
>> MarkMail: http://hippocms-dev.markmail.org
>> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html
>>
>>
> ********************************************
> Hippocms-dev: Hippo CMS development public mailinglist
>
> Searchable archives can be found at:
> MarkMail: http://hippocms-dev.markmail.org
> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html
>
>
********************************************
Hippocms-dev: Hippo CMS development public mailinglist

Searchable archives can be found at:
MarkMail: http://hippocms-dev.markmail.org
Nabble: http://www.nabble.com/Hippo-CMS-f26633.html

Reply via email to