Stack Exchange Network

Stack Exchange network consists of 175 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange
11

Honestly, indexing data of any sort from multiple places in multiple formats is a recipe for disaster. You would be better off setting up your own internal email server, switching to a provider that offers indexing features, or adding a proxy-email server are pretty much your best options. All three of those suggestions come with their own pros and cons. ...


8

Why on earth would you install Google Desktop on a server? That's got bad idea written all over it. I don't mean that question rhetorically, either. I literally cannot think of a single legitimate reason to install it on a server and I'm curious about why you think your situation dictates its usage. If the Microsoft-blessed Windows Search 4.0 isn't doing it'...


6

In reality, ALTER TABLE ... DISABLE KEYS and ALTER TABLE ... ENABLE KEYS do not work !!!! I addressed this back in February 13, 2011. I did some additional digging and found out from InnoDB's mother company, InnoBase Oy (before being made 7 of 9 in the Oracle Borg), that this is indeed the case. This link suggests doing DISABLE KEYS and disabling foreign ...


6

The Fine Manual (which is well worth Reading) suggests that the effect of that statement lasts until a corresponding ALTER TABLE ... ENABLE KEYS statement is executed.


4

I've discovered that the problem is because we are using DFS to access our files on the client. We map network drives to domain DFS roots, and when attempting to search using the mapped drives, it yields no results. If I open an explorer window to the UNC path using the DFS path (e.g. \\DomainDfsRoot\share) then again, it yields no results. If instead, I ...


3

Your .htaccess file cannot magically distinguish "real" users from "bot" users. Since from the webserver's perspective, there is no distinction. However, as a general rule, bots will respect the contents of robots.txt, while web browsers do not. Alternately, if you had some way of determining what was a bot and what was not, you could work that rule into ...


3

Only indexes that were created with {background: true} originally (and hence have that option set in the system.indexes collection) are built in the background once restored. Once you have the database dumped you will have your data in the name.bson file and the metadata (indexes to build) in the name.metadata.json file. Unlike the BSON file, the JSON ...


3

"pinning" in Oracle parlance basically means "ensure that this is always kept in memory". It is frequently used for hot indexes that see a lot of queries against them, but might ordinarily age out of Oracle's cache (and thus have to be re-read from the disk). By pinning the index in memory the index scan will theoretically never require a disk access. ...


3

Depending on what files you are trying to index, you probably need the appropriate iFilters so Windows Search can go in and actually sift through the binary contents of each file and grab out the text so it can index it. After the iFilters are installed, go to Control Panel -> Indexing Options Click Advanced button Click Rebuild PDF iFilters Adobe -- ...


2

Their database may indeed all in RAM, meaning that it exists in a state that does not correspond with any persistent data on the drive. For a simple example, when an application does a SELECT * FROM on your database, that application holds the entire database in RAM. It's easy to visualize in Java: you've just created a great, big array of Java objects. ...


2

You may try solr. It's not an end to end tool, but more like a framework. But it's open source and it's made to index almost everything (including content), optimized for quick and flexible search.


2

There's huge performance improvements from mysql 5.0.x to 5.5.x. Check out these benchmarks http://vbtechsupport.com/606/ and http://vbtechsupport.com/657/. Most of the focus is on InnoDB storage engine performance. If you're still using MyISAM storage engine then check out MariaDB 5.2.5 and higher as it's the only MySQL performance fork alternative which ...


2

Any errors from the search service in the application event log? The "search contents" option appears in a couple places. Within Folder Options, and also the Search Service control panel. In the control panel, each file type has it's own option to control content search. Contents of PDF are not searched by default, you'll need to turn that on. I think it was ...


2

Placing the directory outside of your document root and creating an Alias would get you what you want. That is, assuming your document root is something like /var/www/html, you would create /var/www/private and then add (to your global -- sorry -- configuration): Alias /private/ /var/www/private/ This directory would not show up in a list of files in / (...


2

Physical database size will remain the same but white space will be reclaimed, shown in the early am clean up event application logs. That would prevent the physical database from growing any larger in theory. I've disabled the index before in this same situation (for a few days). The only thing I recall users experiencing is that they get a message ...


2

There are volumes and volumes of information out there that answer these questions. The answer is: it depends. Regular index maintenance and statistics updates are crucial, that much is certain. You have to do some research and tailoring of how and when to do it based on your environment, though. As a good start, read these: Rebuild or Reorganize: SQL ...


2

I think you may be looking for a service or appliance such as the Barracuda Message Archiver. I use the appliance version and it indexes all mail as it is sent / received.


1

This: Set objConnection = CreateObject("ADODB.Connection") Set objRecordSet = CreateObject("ADODB.Recordset") objConnection.Open "Provider=Search.CollatorDSO;Extended Properties='Application=Windows';" objRecordSet.Open "SELECT System.ItemName, System.DateModified FROM SYSTEMINDEX WHERE DIRECTORY='file:C:/usr'", objConnection Do Until objRecordset.EOF ...


1

Windows search on the client can use the index created from search on the server. For pdfs you will need a pdf ifilter. Here is the kb article about how to install the indexing service on the server http://support.microsoft.com/kb/954822/ You could also take a look at MS Search Server Express.


1

The "normal" location is used mostly by PHP-powered and static websites. Django is a whole different beast. In a typical Django setup any file you upload to the web server will be ignored (i.e. it will not be statically served) no matter where you upload it. Since you apparently don't understand how Django works, you'd better wait for your friend to come ...


1

Nowhere in the question does it actually state it must be free (does it?) In that case, the suggestion I have is the Google Search Appliance. It seems to support a whole bunch of windows-type stuff, a massive bunch of filetypes, and sharepoint integration. It's probably not cheap, but it is supported, and it's Google. Who knows search better than they do!...


1

Check out SearchBlox. Or you could always roll your own permuatation using Lucene, Solr, Nutch and/or other similar tools.


1

The value of 84 for HA_FT_MAXCHARLEN is defined as a third of HA_FT_MAXBYTELEN (which is 254). This is defined in include/ft_global.h and changing it will mean recompiling MySQL - it is not a run-time variable. The reason it's a third is because in many character sets one character is not equal to one byte. If you were to change HA_FT_MAXBYTELEN and ...


1

First non-clustered indexes are not new in SQL Server 2008. They have been around for ever. Filtered indexes are new in SQL Server 2008. There is no way to have the system figure out what indexes need to be created before any queries are run against the SQL Server. There are two ways to create indexes in SQL Server. One is with T/SQL, and the other is ...


1

There are so many solutions out there for document management; are you sure you want to reinvent the wheel? Personally, I'd recommend Knowledge Tree (there's a free community version). If you've got your heart set on web page full of file links, you probably want to do it web-side, so it's dynamic and you don't have to remember to re-run it when you put ...


1

Sounds like issues with the index statistics. If you are using MyISAM, sudden spikes of INSERTs would throw the statistics off in the eyes of the MySQL Query Optimizer. This would cause the MySQL Query Optimizer to take very bad guesses within the EXPLAIN plans of SELECT queries. If you are using InnoDB, ANALYZE TABLE becomes completely useless. As long ...


1

Based on this article, it looks like you could reset the accidental ID to a proper value, or drop the row and then restart the server as the auto-increment value for innodb tables is held in memory. I haven't tried this before but it may work.


1

9M rows is not really that many. My guess is that in both situations, something is locking the table. Check the table locks, and either close any processes that are locking the table, or kill the locks manually (if you're sure nobody else is using the database). There's a reasonably good script to show all locks here. SELECT L.request_session_id AS SPID, ...


1

The trouble with storing a database in RAM is that RAM has a nasty habit of forgetting everything when the power goes away i.e. it's not persistent. That said, making proper use of memory for high-traffic sites is absolutely essential to getting decent performance, because you get extremely good IO rates from it and that's a very useful thing if you have a ...


1

Yes, although I think twitter use a variety of technologies and not just RDBMS. There are engines for MySQL that ONLY run in memory for example (the cluster NDB if memory serves). Frequently, yes. Not by definition, but yes some can be. It's often best for any database to maximise use of RAM and minimise slow disk access. Memcached is certainly 1 common ...


Only top voted, non community-wiki answers of a minimum length are eligible