|
|
|
Welcome to our on-line Frequently Asked Question area. This is a collection of answers to the most commonly asked questions about Phantom. If you have a specific question, it is probably answered here, but if it's not, please e-mail us. Even if you aren't looking for a specific answer, there are lots of good tips and ideas here...
Have you looked in the User Guides? Default Phantom URL on Windows.The default administration URL for Phantom on a Windows computer is: [ Back to Listing ] Why is there a lock.txt file in my Phantom folder?This file is used by Phantom to make sure there are no other instances of Phantom attempting to use the same data file. Please do not remove or alter this file. It will only be present when a Phantom instance is accessing that data file. Phantom running as a service will also create a locking file for the data file it is using. [ Back to Listing ] How can I tell if Phantom is running as a Service?The easiest way is to connect to the Web Admin interface. You may also wish to examine the servicelog.txt file in the “Logs” folder of the Phantom install. [ Back to Listing ] I've started up Phantom as a service and I can't connect to it from the Web Admin interface what should I do?The first thing you should check for is a file in the c:\ directory of your computer called “PhantomServiceQuit.txt”. The contents of that file should give you a good idea of what caused the service to quit. If there is no “PhantomServiceQuit.txt” file then the problem may be related to Phantom preferences having been changed. At this point you should go to the Windows Service Manager and ensure that the Phantom Service has been stopped, then start up Phantom as a stand alone application (not a service) and continue with the troubleshooting. Many times this problem can occur when the default Phantom port has been changed to something that is conflicting with another application. In that case it is necessary to shutdown the other application at least temporarily while you change Phantom’s default port number. [ Back to Listing ] What are the files PDFDuppy and WordDuppy doing in my Phantom folder?PDFDuppy and WordDuppy are used to assist Phantom in crawling PDF and Word Documents. These two files must remain in their original install location for Phantom to properly crawl PDF and Word documents. [ Back to Listing ] Why is there a in.pdf, in.doc, out.txt, or done.txt file in the Phantom folder?These files are used by Phantom when crawling PDF and Word Documents. While performing a session update these files may periodically appear in the Phantom folder. This is normal behaviour. [ Back to Listing ] Phantom crashes on my machine. What should I do?Phantom is generally quite stable, but here is a list of basic things to try if you experience crashes when running Phantom. On Windows 95 or 98, make sure "Disable the Phantom cache while indexing" is checked on the Robot Preferences page in Phantom. This is necessary to work around memory problems in Windows 95/98, and will significantly impact performance. For this reason Windows 95/98 is not recommended for running Phantom. If the machine has previously crashed or been interrupted (by a power failure, forced reset, etc.) while Phantom was running, the data file may be corrupted. This can also happen if Phantom is force quit using cmd-option-esc on a Macintosh or the Windows NT Task Manager. Run 4D Tools to rebuild your data file. See the FAQ Fixing Corrupt Data Files with 4D Tools. On Windows NT, make sure "Maximize throughput for Network Applications" is selected in the Memory control panel. On Macintosh, increase the amount of memory allocated to Phantom by 1MB or more. Phantom should have a minimum RAM allocation equal to 10MB plus 10% of the data file size. For example, if the data file is 26MB, Phantom should have at least 12.6MB allocated. 10MB plus 2.6MB (10% of 26). For Windows, applications set their own memory partitions dynamically. Make sure you have free memory equal to at least 10MB plus 10% of the data file size when Phantom is not running. Note that this must be physical RAM: increasing the swap file size when there is insufficient memory will only decrease performance and increase the frequency of problems. If it has been a while or if you have installed OS updates on the Mac running Phantom, try reseting the PRAM (hold down Cmd-Opt-P-R keys at startup waiting for the system to chime 3 times). You will have to re-set some control panel settings, including AppleTalk, after resetting the PRAM. Update your disk drivers when installing a new Mac OS version. Older disk drivers are often the problem with Type 11 errors. On Macintosh, remove any extensions that are not part of the standard System Software installation, especially any Microsoft extensions. (Applications like Microsoft Word or Excel really have no place on a Web server so you should probably remove them.) Performing a clean install of the Mac OS is the surest way to eliminate any non-standard extensions. This is also a good idea if you have not done one in a while. There are known crashing problems using Dayna BlueStreak PCI ethernet cards with Macintosh Performas, Power Macintoshes and Mac OS compatibles having PCI slots but that do not have built-in ethernet. [ Back to Listing ] How can I tell Phantom to use a different data file?To change your choice of data files on Macintosh, hold down the option key while Phantom is starting. On Windows, hold down the Alt key. You will be asked for the name of the data file to open. Phantom will open the last data file it opened or created each time it is restarted unless that file is no longer there. [ Back to Listing ] Why does Phantom re-index all of the files when I update a session?Phantom depends on the Last-modified date field in the HTTP document header to decide if it should re-index a document when a session is updated. That field isn't required, so not all servers provide it; some servers or plug-ins (namely SSI and NetCloak) may also suppress the last-modified date. Browsers use the last-modified information to decide if they should re-load a document they have already cached. They do that by sending an HTTP request including an If-modified-since: field with the last-modified date. The server is then supposed to return either the document if it has been modiied since the date in the request, or an HTTP header with a Not-modified code Phantom will re-index any document with no modified date on an update. NetCloak 2.5 and later has an option to return the last-modified date information so that Phantom won't have to completely reindex a cloaked site. Some servers and SSI plug-ins do not return a last-modified date; see the documentation for your server or plug-in to see if there is an option to return last-modified dates. [ Back to Listing ] Phantom says that my data file is corrupt. What can I do to fix it?Data files can become corrupted for a wide variety of reasons, but corruption most frequently occurs when the machine crashes or is reset while Phantom is updating its index. To prevent this, avoid resetting the machine or forcing the Phantom application to quit or exit using the Task Manager (Windows NT) or command-option-escape (Macintosh) unless it is absolutely necessary. Sometimes the data file will be unrepairable, but frequently 4D Tools can repair a corrupted data file. You can download 4D Tools right here. To Recover:
Rebuilding the data file will create a new file which includes as much data as possible from the old file. It is unusual for data to be lost in the recovery process. Normally if the data file is damaged badly enough so that significant data will be lost it will be unrecoverable or the rebuilt file will be unusable. When you start Phantom again, remember to hold down the option key (for Macintosh) or the alt key (for Windows) while Phantom is starting up so that you can select the new, recovered data file. Note that there may be problems with the data file that recovering with 4D Tools will not fix. If you are having problems (especially crashing or freezing) that persist after recovering the data file as described here, or if 4D Tools crashes or freezes while attempting to recover the data, you should create a new data file and re-index the pages. If you have already used 4D Tools to Recover by Tags, the recovered file will already be compacted so compacting it will have no additional benefit. Note that running 4D Tools to either recover compact the data file may take a long time- usually 2 to 5 minutes for each MB of data in the file, depending on the speed of the machine where it is running. For example, compacting or recovering a 20MB data file will normally take about an hour or two; a 200MB data file will take between six and eighteen hours. If you decide to run 4D Tools to recover a copy of your data file, quit or exit Phantom before making the copy, or use a backup from the Backups folder. With very large data files it is often more effective to restore a backup than to recover the current data file. [ Back to Listing ] What do the Page Result Codes mean in Phantom?The Page Result Code in the Page Browser indicates what happened when Phantom tried to index the page. All of the 200-500 series are HTTP response codes, defined by the HTTP spec. The 1000 series are specific to Phantom and are related to what it got when it tried to get a particular page. The -1 through -4 indicate problems when Phantom tried make TCP connections and failed. The other negative numbers are system error codes (like -35 for volume not found). 200 OK201 Created 301 Moved Permanently 302 Moved Temporarily 400 Bad Request 401 Unathorized 403 Forbidden 404 Not Found 500 Internal Server Error 502 Bad Gateway 503 Service Unavailable 1001 URL is outside valid directory 1002 URL is in directory excluded by robot exclusion standard 1003 URL did not pass filter 1004 Empty stream - no data returned 1005 Timeout occured while receiving file 1006 URL removed from session and index -1 Bad sessionID passed to TCP Toolkit routine -2 TCP Toolkit timeout has expired for this operation -3 Internal error, contact AKTIV Software -4 Negative number of bytes requested by TCP_RecvNBytes -35 Volume does not exist -37 Bad file or volume name -43 File not found -120 Folder not found A complete list of page result codes is included in the User Guide under Sessions. [ Back to Listing ] How can I move a Phantom data file from the Mac to the Windows version?Data files on the two platforms are almost identical, so you can transfer them from one platform to the other with no problems. You'll need an utility called 4D Transporter to convert the file. It only runs on MacOS, so you will have to run it either before or after you transfer the file depending on whether you are moving the data file from Mac to Windows or vice-versa. Depending on the version of 4D Transporter being used, a .4DR file may be produced, or required, by the conversion. [ Back to Listing ] How can I change the HEIGHT and WIDTH attributes that Phantom applies to the New Page icon?In Phantom 2.2 and later you can edit the "New Page GIF HTML" template on the Result Data page of the session or preferences. In Phantom 2.0 and 2.1 these are stored in a resource, so you will need ResEdit to change them. Open a copy of the Phantom application in ResEdit and edit 'STR#' resource ID 16004, string 7. You can change the height and width, or remove them by changing the string to a space. Editing resources is only possible on Phantom for Macintosh. We plan to add the settings that are currently in resources to the Admin interface, but until then, the height and width cannot be changed with Phantom for Windows 2.1 and earlier. [ Back to Listing ] Can I format a Phantom search request as a link instead of a form?Yes! Phantom supports search requests by either POST or GET methods. To put a search request into a link, include the normal search form fields in the URL after the Phantom CGI name separated by ampersand ("&") characters. You must specify the information like search text, max hits, etc. in the link, of course. Link searches are useful any time users will need to perform the exact same search. Putting the search into a link means that users won't be able to enter any information or select any options, like detailed results or search text, but it means that performing the search becomes just a matter of clicking the link. For example, you could make a "find what's new" search for your site. The search form might look like this: Find what's new on our site! Show and show pages that are new or updated since the <INPUT TYPE=submit VALUE="Search"> Users get to select the number of results and the length of time for the updates, then click the Search button to perform the search. Here- try it out! You could make this a link, so all the user has to do is click on it to get the results. The HTML for the link would look like this: <A HREF="/Phantom.acgi$search?.Session4=4&.maxHits=25&.lastUpdate=day"> Find what's new on our site since yesterday!</A> This link will always find what's new in the past day and return 25 results. You could easily add additional search links for the past week or month, too. <A HREF="/Phantom.acgi$search?.Session4=4&.maxHits=25&.lastUpdate=month">Find what's new on our site since last month!</A> [ Back to Listing ] How can I access Phantom if I forgot my Remote Admin password?There are two ways you can reset the username and password for Remote Administration. First, you can access Phantom from a Web browser runnning on the same machine as Phantom without using the username and password. Access the Admin Home Page with the usual URL, then go to the Preferences page and enter a new username and password. Second, the Remote Admin username and password are reset when you install a new copy of the Phantom application. (The default username is "Phantom" and the default password is "Duppies".) You can install Phantom into another folder on the same machine where it is running, then copy the Phantom application (Phantom.exe for Windows) into your existing Phantom folder. Once a password has been entered, there is no way to find out what it was if you forget. Just choose a new password and enter it. [ Back to Listing ] What should I do when the Robot won't start?If crawls are not starting like they should, Phantom may not have enough memory. This may be either when clicking the "Run" or "Update" links on the Admin page, or when an update is scheduled with the timer. For Macintosh, quit Phantom and then increase the memory partition by using the "Get Info" command in the Finder. The default memory partition is enough to crawl a few hundred pages. If you are crawling more pages than that, more memory will be necessary. If there is insufficient free memory to increase Phantom's memory partition, quit unused applications and remove unnecessary system extensions. Increasing the amount of virtual memory is not recommended, as this will have a significant negative impact on Phantom's performance. For Windows, exit or remove applications or services that are not necessary. Increasing the size of the Windows swap file is not recommended, as this will have a significant negative impact on Phantom's performance. A 16MB system may not have enough memory to run Phantom, even if you quit all other applications. [ Back to Listing ] How much RAM should I allocate to Phantom?On a Macintosh, you can set the application's memory partition using the Get Info window in the Finder. Phantom should have at least 10000k plus 10 percent of the size of the data file. For example, if your data file is 48MB, you should set Phantom's memory partition to at least 10000 + 4800 = 14800k. If you have the number of listeners set to more than ten you should add another 200k for each listenter above ten. These settings apply when Virtual Memory is on. We strongly recommend that you run Phantom with VM turned on, and the virtual memory size set to 1MB more than the amound of physical RAM in the machine. If you can give Phantom more memory it will use more until the application's RAM partition is as large as the data file. Phantom will use the additional memory to cache the data from the data file, resulting in faster searches. Allocating more won't hurt anything, but will not have any additional benefit. Under Windows 95, 98 or NT, Windows will automatically determine a RAM partition for an application, and there is no user setting. The most common symptom of not enough RAM is that the robot will not start when a timer is scheduled or when you click the "Run" or "Update" link on the Admin Home page. If you have this problem on a Macintosh, allocate more RAM to Phantom. On either platform, you can also reduce the number of listeners in Phantom, quit other applications and/or add more RAM to the machine. [ Back to Listing ] What is Phantom doing when it displays "Removing Index" in the status window?Phantom is removing pages that have been deleted from the index. When a page is deleted (either through the page browser or on an update), it is not actually removed from the index but rather is marked as deleted. The deleted pages are actually removed later by a background process; this is what you are seeing. Removing deleted pages can be time consuming, which is why it is done by a background process. Cleaning the index is important, though- Phantom will be slow and erratic if there are too many deleted pages that have not been removed. If you have deleted a large number of pages or an entire session, removing the pages may take several hours or more. After the keyword cleaner has finished removing deleted pages you may want to run 4D Tools to compact the data file. If there have been a large number of pages removed this can significantly improve Phantom's performance as well as decreasing the size of the data file. [ Back to Listing ] Why do I always get 401-Not Authorized result codes when trying to index pages on an MS-IIS 3.0 server? The pages are freely accessible from a Web browser.Some IIS 3.0 servers return a 401-Not Authorized response to requests that have an unrecognized user-agent. Phantom normally sends a user-agent of "Phantom" and its version number, so it sometimes has this problem. A workaround using Phantom for Macintosh is to use a hex editor (like HexEdit) to change the user-agent string returned by Phantom. The string is located at byte 194E4D in the data fork, or, do an ascii search for "User-Agent: Phantom/". Change "Phantom" to "Mozilla" and that should solve the problem. Thanks to Franck Horlaville for discovering this workaround. [ Back to Listing ] How can I return search results with frames?Crawling a site with frames is no problem, but presenting search results and linking back to pages on a framed site is more of a challenge. Just as there are a variety of ways to use frames on a site, there are a variety of ways to present search results and links to pages on a framed site. Phantom cannot return the frameset and all of the frames exactly as they would appear to a user on the site. There could be many combinations of frames and framesets with the same page, so that would be impossible. Similarly, since a single document can appear in many different frames or framesets, Phantom does not store any information about what frameset it found a particular document in. If your site uses frames consistently, though, there are several ways you can return search results in frames. There are a set of examples that illustrate what can be done. First is the easiest way of using Phantom within frames- putting the search page in the body frame. Search results and pages linked from the results will appear in the same frame as the search page. Second is an example of a common mistake when including a search form in the navigation sidebar of a frameset. Third is the same example, but with a TARGET tag so that the navigation sidebar search form correctly directs results to the body frame instead of the sidebar frame. Fourth is similar to the third, but with a twist- the search results appear in one frame, but the pages linked from the results page appear in a different frame. This requires more than just a TARGET tag in the search form. The last example takes the fourth one a step farther. Pages linked from the search results don't just appear in a different frame, but in an entirely different frameset. [ Back to Listing ] Why doesn't my serial number work?When you type a serial number into Phantom 2.5 and click the OK button, Phantom will should accept the number and display a message indicating so. If it does not accept the serial number, the computer will beep. The entry box will remain allowing you to re-enter the number and try again up to five times. If your serial number is rejected, you may have mis-typed it or it may not be valid. If you will be typing the serial number, remember that the serial numbers are case sensitive. Take care to distinguish lower case "l" characters vs. upper case "I" and the digit "1" (one), as well as upper case "O" vs. the digit "0" (zero). If you received your serial number via e-mail, the easiest way to enter the serial number is to copy and paste it directly from the e-mail message. If your e-mail is on a different computer from Phantom you can paste the serial number into a text file or as the name of a folder, then copy that to the server running Phantom. On the server you will need to copy the serial number to the clipboard before launching Phantom. Finally, just paste it with command-V or control-V when the serial number entry box is displayed. You may also check your serial number using Check It! If you have a license for an earlier version of Phantom you can upgrade to Phantom 2.5 for less than the full purchase price. [ Back to Listing ] Can I run Phantom as a Windows NT Service?Yes. As of version 2.2.2 Phantom can run as a Windows NT/2000 service. [ Back to Listing ] Can I make a Sherlock plug-in for Phantom?Yes! You can create a Sherlock plugin to search your site (or other sites indexed by your copy of Phantom) quite easily. Download 'Sherlock and Phantom' package. Sherlock plugins are site-specific, so there is no generic Sherlock plugin which will work with any Phantom-powered search. You must create one for your site, or modify the sample included in the package. But don't worry- it is relatively simple, and instructions are included. You'll need a Macintosh to run Sherlock, but you can make a Sherlock plug-in for your site regardless of whether you run Phantom for Macintosh or Windows NT. [ Back to Listing ] Last updated |
||
|
|||
