Search the Community
Showing results for tags 'performance'.
Found 23 results
-
Environment: FMS 15 Windows Server 2012 host on Azure Plug in: 360Works ScriptMaster Purpose for using the Plug in: zip the pdf files. Problems We have: There are a large amount temp files created on the client machine which run the fmp to connect the server , the temp files won't removed. What I am looking for: Any script step can avoid these temp files stay in temp folder. How these files created? I attached a screenshot. Thanks
-
Hey again. In a previous thread I suggested using javascript in a webviewer to return the difference between the server and client UUID lists in order to increase the sync check speed. It was much faster, but the code I used included an algorithm with quadratic time complexity, meaning that as the size of the lists increased the time that sync check took would grow exponentially. @GisMo helped test confirm that this solution would not scale well. However I think I've found a bit of javascript that solves this issue and will scale linearly. I've changed the script from the previous thread to use this new code instead. If you are using this javascript method for list comparison I highly recommend you change the sync check script to use this code. Simply open the new file, copy the calculation from set variable[ $url ] in "Sync Check Client", and paste it into your Set Variable[ $url ] step. Otherwise if you're using EasySync and struggling with sync check, give my solution a try! I've attached the new file below and linked to the old thread for continuity. Best, Josh EDIT: For anyone interested in a demonstration of just the algorithms' performance difference before updating sync check, I've uploaded a raw concept file. EasySync_JWH_SyncCheckMod_LINEAR.zip Linear_Array_Comparison.fmp12.zip
- 24 replies
-
- 2
-
-
- easysync
- fmeasysync
- (and 2 more)
-
Unstored calculation fields and performance
TJ53 posted a topic in Calculation Engine (Define Fields)
Let’s say we have two related tables: “Invoice” and “Invoice_Item”. We could create a calculation field in the “Invoice” table called “total_amount” with this formula: total_amount = Sum (Invoice_Item::amount) This field would have a negative impact in performance when appearing in the layout, since it would have to be defined as unstored, because it’s referencing a field from a related table. Now let’s suppose this field is not used for any scripts, tooltips, conditional format, etc … would the performance of the database be negatively affected ONLY when this field appeared in a layout? In other words, would adding an unstored calculation field to a table involve a performance penalty, even in the “unreal” case where this field didn’t appear in any layout, script, conditional format, etc.? thanks in advance! -
Is there any difference in terms of performance between a calculation field (stored and indexed) and field defined as auto-enter calculated value (indexed)? For example, we have an “INVOICE” table, with a field called “date_invoice_sent”, and we’d like to have a boolean field called “is_sent”. The calculation would be “not IsEmpty(date_invoice_sent)” So we have two options here: - Calculation field (stored, number result). - Number field defined as “auto-enter / calculated value / do not replace … unchecked”. Would there be any difference in performance between the two options? thanks in advance!
-
I have a system being hosted using FMS14. I also have a duplicate database hosted for testing/exporting using FMS15. PSOS is much slower in my hands with FMS15. Both servers are physical boxes, with similar (maybe identical) specs (Cores, RAM HD Space, HD Space available) I also have a Server Side script that creates a found set, exports data to excel and emails it to a user. (This is based on https://www.skeletonkey.com/restoring-filemaker-clients-found-set-within-server-side-script/). The exports have a couple of related fields, but no unstirred calculations. Using FMS14 this has been working just fine. Using FMS15 it has been significantly slower. As a test, I exported 13,000 records from both FMS14 and FMS15. Using FMS14, it takes about 10 Seconds to recreate the found set, then 70 Seconds to create the export. Using FMS15, it takes about 20 Seconds to recreate the (same) found set, then 175 Seconds to export the data. In another series of tests, this time exporting 30,000 records (more realistic in my scenario) I found; Using FMS14, it takes about 25 Seconds to recreate the found set, then 168 Seconds to create the export. (2.5x longer to find/2x longer to export 3x records) Using FMS15, it takes about 90 Seconds to recreate the (same) found set, then 825 Seconds to export the data. 9x longer to find/11x longer to export 3x records) In the last series of tests, this time exporting 50,000 records I found; Using FMS14, it takes about 59 Seconds to recreate the found set, then 262 Seconds to create the export. (6x longer to find/4x longer to export 5x records) Using FMS15, it takes about 160 Seconds to recreate the (same) found set, then 1700 Seconds to export the data.(16x to find/24x to export 5x records) Any ideas? Anyone see similar results with PSOS on FMS15? Again, the machines are identical, the databases are identical and the scripts are identical. the only difference is FMS14 vs. FMS15. I am also letting another user pull data using ODBC. That has also gotten extremely slow using FMS15, but that is a discussion for another thread. I am not using WebD on this server. In general the FMS15 database performs find using FMP Clients. But the FMS15 functions are not very impressive. I am petrified to move my production database to FMS15, and even considering moving the test server back to FMS14. Help! TIA Jerry
-
Performance of using UUID as a primary key
VictorLion posted a topic in Interface Design Discussions
I am building a customer database. The database will have a primary key. I am looking to make the primary key either a unique integer or a UUID (with 32 chars). The UUID appeals to me due to its ability to sync tables. The primary key will be used for (obviously): Searching and sorting records As a tag against documents eg document XX is linked to customer AA, BB and CC; and A foreign key to link tables My questions are: Do UUIDs adversely affect database size Do UUIDs adversely affect speed If UUIDs retain their full 32 chars, then there will obviously be a disadvantage. However, is FM clever enough to convert the UUID "under the hood" (such as to an integer referenced index) to improve database speed and/or size. -
Hosted Filemaker Server Performance- International access
cat traveller posted a topic in FileMaker WebDirect
Hi everyone, currently I am looking into creating a very simple database for a cocktail competition for a spirits brand. The competitors are coming from markets all over the worlds, the local offices are submitting the recipies, images and some further information in my ideal world via webdirect into the database itself. I dont think there need to be more then 2-3 users accessing that database simultaneously. My question would be if anyone is experienced with filemaker hosters and the perfomance of that hosting when the db is accessed from different parts of the world. In other words: where would be an ideal country for hosting the file (in terms of performance) if it were to be accessed from Europe, Japan, USA, Israel, China and South America? Please advise, Thank you. -
Slow layouts become faster after changing Auto-Enter calculation
MoMo posted a topic in FileMaker Server 13
Hi all My server is running on FMS 13.0v9 and my clients are running FMP 12.0v4. The mismatch is due to my clients still mostly being on OS X 10.6.8. I'm having a strange issue with slowness that I was hoping to get some insight on. Here goes: I have a layout (named "Checkin Layout") that access a table and a few related tables. The main table has a field named "CheckedIn" which contains either a 1 or a 0. The user checks a checkbox, which enters the value "1". The auto-enter calculation is as follows: Case (CheckedIn = 1 ; 1 ; 0) This makes it so that the value isn't ever empty, but will be either 0 or 1. Users have lately been complaining that toggling this checkbox is becoming incredibly slow. I initially tried to remove objects from the Checkin Layout, but found that even if the layout contained nothing but the field in question, toggling the checkbox was still incredibly slow. I then tried changing the auto-enter calculation, thinking that may help. I changed the "Case" function to "If" and noticed that the speed shot up immediately. However, I found that if i closed and reopened my layout, the slowness came back. After some more experimentation, I found that opening the "Manage -> Database" window and changing ANY definitions or calculations would make the Checkin Layout extremely fast, until it is closed. Then once it is opened it becomes slow again. I'm curious if there is anything happening when I adjust definitions that I could do via scripting. The slowness is taking its toll on productivity, and I can't be around to mess with definitions every time a user needs to check something in. Is there anything I can do?? Thanks in advance for your help. MoMo -
Hey Fellow Easy Sync-ers. I have an idea for improving the speed of the Sync Check step. Instead of comparing UUID lists in loops, we can retrieve a UUID list from server, and compare it to the client UUIDs using a couple lines of JavaScript in a webviewer. I've tested this and my Sync check time comparing ~23000 records, over 18 sync tables on a shared server has gone from 5-10 minutes down to a few seconds. No plugins necessary. Process: Get UUIDs_client. Get full UUIDs_host list from server Format both as JavaScript arrays and put them into a web viewer to be compared using 'filter' Have the webviewer call another script using FMP url to return the resulting diff array as a list. Process the diff like normal (The way Tim wrote it). I also updated the Sync Check process to pull the Server UUIDs in segments if they exceed the $$max_pull_segment_size setting. Here are my example files. Try adding a bunch of records and then run the sync to see how quick the sync check is. PLEASE NOTE: This method requires a named webviewer on the client EasySync_Payloads layout. EasySync v1.3 mod1 (jwillinghalpern@gmail).zip
-
Understanding and tuning Filemaker Server Performance
Nick Lightbody posted a topic in Community Articles, Tips, & Techniques
In my view, in the world of FileMaker Server the concerns of owners and admin with FMS11 and earlier are little different to those of FMS14 admins today. Yes, the product has developed and we now have the new appearance layer from 12 onwards, and WebDirect from 13 on. But, the underlying issue remains, how do you get to understand FileMaker Server and get the best out of it, when it is trivial to build a very "bad" FM solution that is slow and unscaleable and when FMI provides no comprehensive and clear advice? FM software does little to prevent folk using the freedom of FM in that manner, because as an "enabling" and "accessible" type of software it lacks the controls, the digital police, that a more prescriptive system may include. It also works very hard to keep users of legacy systems on-board and supported and their solutions inevitably include much that would regarded as poor design today and hence cannot easily deprecate old techniques, so FMI's options are limited. Hence my having been developing a "mobile first" design framework since April 2012, and having eventually worked out what does work, I was encouraged by various folk to develop my report on "Understanding and Tuning FileMaker Server Performance" and publish. I have also published an early version of part of the report here on FMHacks and done two videos for Richard Carlton Consulting here and here which cover some of the report material. My work has been validated by folk at FMI (who wish to remain anonymous), Wim Decorte, Richard Carlton and Alan Stirling. This is the first detailed and comprehensive report on this subject for many years, you can read it here and download a pdf here. Many of the techniques I have developed, and to which the report alludes, are included in the free and open source Deskspace dsBenchmark tool, which can be downloaded from here. I would encourage your responses, and disagreement, if you feel so moved. We all move forward by rigorously testing both the old and the new and no doubt you can offer improvements and additions to what I have constructed. Where I am persuaded, your input will be included in the ongoing Report and duly credited. The current version of my Report is of course the html pages on deskspace.com, not a dead paper document. In fact wasn't that what Sir Tim Berners-Lee had in mind when he invented and donated http, html and the World Wide Web? This ease of editing is facilitated by our new Deskspace rwApp creating the responsive web pages that comprise www.deskspace.com - but that is perhaps another story. Cheers, Nick -
Import Script performance from Windows is very slow
haseebiqbal posted a topic in Importing & Exporting
Hi My scenario is, when I used to RUN a Import (Folder with Import only the reference) script from a iMac mini to remote Windows File server, it used to take 4-8 hours. But when started using the same script from Same server(Windows), it started to take more than 24 hours. Only things I have changed: Import from Windows FS ( Files of type Windows from OS X) Modified the Folder location to local server. Any suggestions or known issues.. Pls note: Previously we used Filemaker 10-11 now using 13-14 in both client server. The old Mac was OS X (10.5.8) with FM Advanced 10. And the old remote server was a Physical windows 2008 server with FM 10 server running as well as Fileserver to hold the files which where been imported as reference. -
I have a vertical market application that uses a php modale that gives mobile users access to their agenda (FM11). Currently I am rewriting this application from fp7 to fmp12. As part of this upgrade I want to offer mobile users an off-line module on iPads. Mainly because mobile users complain that they do not always have a good internet connection when on the road, so the performance is not always acceptable. I have implemented FM EAsySync in my application on a test server and on a small scale this looks good (Tim awesome job in setting up this framework). However, on this forum I noticed some comments about the processing time of synced records, so I am wondering if anyone has already succesfully implemented and rolled out this framework in an application. If so, I would like to hear about your experience in using this framework. The last thing I want is to offer an off-line module to my customers as an improvement over the current php module (and require my users to purchase iPads) and then find out that there are other performance problems. I am not asking you to comment on the performance in my application. I realise this can not be done without detailed information about the application itself. I am just curious to find out who has already implemented this framework and rolled out an application that uses this framework. Thanks, Joost Kingma
-
Comparison between layout load times and performance
Stu412 posted a topic in Calculation Engine (Define Fields)
Hi there I am experimenting on a sample of data in a very simple FM database. I need some help on 'how' FM does things, rather than how to do something myself (although you may scream at my technique below!) I have two customers on Customer_Table with CustomerID and Name as the fields. The only other table is called Data and is joined to Customer via CustomerID_FK. It's a one to many relationship as each customer can have many data rows, around 400 each. Therefore I have around 800 records in the Data table. I have added three text fields to the data table which enable true/false conditions to be set. I use many CASE statements on the Data table to select various values from records depending on the combination of true/false conditions. I then summarise these on a second field to get the grand total per true/false combination. An example of this would be: CustID Text1 Text2 Text3 Amount CaseStatement1 CaseStatement2 10 North Confirmed Bob 11000 11000 10 North Possible Bob 5000 5000 If I want a grand total of Bob's confirmed and possible sales in the north, I'd do: Case ( Text1 = "North" and Text2 = "Confirmed" and Text3 = "Bob" ; Amount) Case ( Text1 = "North" and Text2 = "Possible" and Text3 = "Bob" ; Amount) I then create summary fields for CaseStatement1 and CaseStatement2 to get the specific values I want. As I say, given the details in the data I have and the requirements of the task, there are many similar calculations through my test DB. The next part of my test was to see how I can easily move between customer records to view which sales have happened, ie, view the summary fields. I set up two layouts, one based on Data and one based on Customer. Given the data I have and the volume of summaries, I noticed it was significantly quicker moving on the Customer layout between records 1 and 2 than using the Data layout (which of course has 800 records). What I don't understand is that surely FM still needs to calculate the Case and Summary statements while moving between just the two customer records regardless, so why is it so much quicker one way versus the other? Thanks in advance -
Building on the success of classes offered for the first time in Europe last year, an announcement has gone out today about a new series of three-day FileMaker Master Classes offered from March to June 2015 in: • San Francisco (March 2015)• Berlin (April 2015)• London (April 2015• Westport CT (May 2015)• Sydney (May 2015)• Tokyo (June 2015) The classes are unashamedly targeted at advanced developers, and are packed with high level content ranging from designing for scalability, advanced development techniques, process control, FileMaker application and deployment models, solution design perspectives to code logic. From the 'brochure': "These intensive three-day seminars are dedicated to expanding your knowledge and broadening your approach to FileMaker development." Full details including event dates, pricing, locations, detailed content outlines and online registration can be found at: http://www.nightwingenterprises.com/MasterClasses The deadline for registrations for the first four locations (San Francisco, Berlin, London and Westport) is 23 February. It would be great to meet some of you at one of the events in this series! Regards,Ray
-
Hi Everyone, I have a database table containing around 700 fileds and half of them are auto enter calculations. Every time i create a new record it takes a long time to create, even more the first time. I just need the record to be created, rest calculations can be done afterwards through scripting. I cannot remove the calculations. Have tried creating the record in seperate table(with the required data only) and then importing those data into my main table, but still, it takes a while. So i was thinking of a way if i can supress the calculations while the new record is created. Is there any or other way of achieving that? In desperate need of this performance tuning. Thanks -
Performance: ExecuteSQL() versus List()
Alex Taylor posted a topic in FileMaker Query Language or FQL
I've been working on optimizing a complex part of a solution that relied on the List() function to send multiple parameters to a custom function. Through my research I came across this article: http://www.teamdf.com/weetbicks/the-search-for-fast-aggregates--trial--error/99/ This article discusses the fact that when you use List(), Sum() or other aggregate functions, FileMaker sends the entire record set to the client, not just the data from the requested field(s). Does anyone know if this is also the case with ExecuteSQL()? I would assume that an SQL query would only return the requested data, but I'd like to know if this is definitely happening or not. -
All, I am building a solution which requires ESS connection to Oracle 11g from FM12. Got the ActualTech Oracle ODBC Oracle Driver. Setup went fine. Managed to do imports from Oracle using the Import Script Step. Can display ESS Shadow tables from Oracle in the FM Relationship graphs. However: * when doing so, FM tells me that the External Table "does not provide a pre-defined unique key" and asks me to choose a field or field combination to identify the same * All relationships i perform involving an ESS table display the 'T' cardinality symbol implying that the match field has no index information. The subsequent problem is obvious : performance as we are reading hundred of thousands of records. It looks like the oracle index information cannot be read whilst it does exist. I have checked with the Oracle DBA : the Indexes do exist on the table (not view) / fields we are accessing via ESS and the Oracle user used for ODBC connectivity does have adequate permissions to read both all_indexes and dba_indexes tables in Oracle. This has been confirmed using an straight SQL query in Oracle SQL developer. Hence my question : does any one have an answer to the following questions : * Is the issue related to Oracle ? ODBC technology itself ? Actual Tech ODBC Driver for Oracle ? ESS Technology in FM ? Setup of any of these layers ? * Anything the Oracle DBA should be doing further so that FM can actually read the existing Index information ? * Anything I am obviously forgetting ? Any hint welcome Thanks Laurent
-
Hi, some time ago, I posted this technique called FMSDIFM for FileMaker Server, Do It For Me. Although it was widely discussed and used by FileMaker 'techies', it was too hard to deploy, so we had to keep it for ourselves (the principle was already public but the implementation was only for our hosting services customers). Since then, we've worked on a simplified version so every one can install it on any FileMaker Server hosted file. We've been using it for a year, and now we're making it public and free. I will present a session at Pause On Error Berlin this week about it. If you're coming to Berlin, bring some heavy script of yours that you've been trying to optimize, and I will show you how FMSDIFM can help. Here is a new video and the download link. Please consider the web page down the link to be the 'official' documentation page, and post your comments here on FM Forums. The-page-with-a-video-and-everything. For those who don't know yet what FMSDIFM is about, well, it's the same thing, read the page Seriously, FMSDIFM is about performing CPU/data-intensive tasks on the server rather than on the client. It improves performance by a factor of several hundreds. Simply hudge! But as you will see, this is only one of the benefits.
-
Hi, I just posted a new entry on 1-more-thing blog, presenting a 'revolutionary' technique allowed by FileMaker 12. It completely changes my life as a developer, and makes it possible to create very efficient online solutions using FileMaker Pro/Go. It's all explained in this video. -
Last September I wrote an article about a custom function that I optimized to evaluate hundreds times faster. At the end of the article, I challenged my readers and myself by claiming that the already optimized custom function can be optimized even further. Do you remember? Later on I actually really optimized it again, and talked about this optimization during my session at Pause On Error [x] London 2011. Now you can watch the video of this part of my session below: Read more and download my updated sample file at honza.24uSoftware.com.
-
24U FM Bench Detective 1.0.2 Released
HOnza posted a topic in FileMaker Product & Service Announcements
Prague, CZ -- December 5, 2011 -- 24U Software announces an immediate availability of 24U FM Bench Detective 1.0.2 for FileMaker Pro 11. 24U FM Bench Detective is an open solution designed to help FileMaker developers easily add precise benchmarking to their solution. With the help of this tool developers can measure performance of their scripts, calculations, and other parts of their solutions, and optimize them for better performance. FM Bench Detective is a free part of the new product being developed by 24U to address the need for optimizing performance of not only FileMaker solutions but also companies using them. What's new in version 1.0.2: - Internal global variables renamed to prevent naming conflicts - Timestamps are now logged in the host's time rather then client's time to ensure consistency - 24U Toolbox Plug-In 1.1.2 now correctly returns local time instead of always returning GMT+2 24U FM Bench Detective lets FileMaker developers: - Precisely measure execution time of their scripts and calculations - Quickly add measuring code to any existing solution - Store all measurements in a simple FileMaker database for further processing - Differentiate collected data by username, IP address, or software/hardware configuration - Use a single point to control benchmarking level or completely disable benchmarking Use 24U FM Bench Detective to better understand your code There is no magical solution that will take your code and optimize it. It's still you, the developer, who's responsible for the quality and performance of your solutions. But it's difficult to intelligently decide how to implement whatever feature you are thinking of just now if you don't know what to expect from the technology you're using. 24U FM Bench Detective lets you easily measure your code execution time with microsecond precision, so that you exactly understand which operations take longer to process. Only that way you can base your decisions on true knowledge instead of guessing. Since the first version of 24U FM Bench Detective has been released, many FileMake developers started to optimize their solutions. Steve Sanders commented on the 24U FM Bench Detective's website: "I keep track of my Custom Functions in a database. When a Custom Function is revised, I compare some “Sample In” results to previous “Sample out” results to see if I have caused an error during the revision. I also want to compare the execute times to the previous revision. That lets me know if I have optimized or slowed down the Custom Function. I used to use “Extra Suites” on the Mac for this but it was very cumbersome as I had to find and remove my own overhead. I could only get an approximate overhead value. FM Bench will be ideal to re-implement this in a clean manner." Serious developers have always cared about the performance of their solutions. But it took them too much effort to actually measure the performance of everything so they did that mostly only when issues appeared. 24U FM Bench Detective makes developers' dreams come true. Now you can base your decision on exact measurements and it does not take more than just a few clicks to add the measuring code to your solution. 24U FM Bench Detective 1.0.2 is immediately available for download from the 24U FM Bench website upon free registration. More info: http://FMBench.com/detective This announcement is intentionally brief to not bother members of this mailing list who are not interested in detailed information. To read complete Press Releases, please visit the following page: http://www.24usoftware.com/pr 24U Software offers solutions for people, not for computers! -- HOnza Koudelka Software Division Manager, 24U s.r.o. FileMaker 8 Certified Developer FileMaker Business Alliance Member mailto:sales@24usoftware.com http://www.24usoftware.com ### -
What is the optimal setup for hard drives, RAM, processors to maximize performance for a growing database running on FM Server 11 and FM Pro 11? We will have about 20 onsite clients, 10 VPN clients and 5-10 iPad clients maximally for the next few years. We are sticking with Mac and looking at either a top end mac mini server or investing bigger in a mac pro tower. We are running on a mac pro tower purchased 2 years ago and seeing serious dragging during busy work days. we have 2.66 GHz Quadcore with 6 GB ram. solid state hard drives worth it? how much RAM is too much? is it worth buying the top end 12 core "westmere" processors? We can afford to go there but I would prefer to use our money on database customization and enhancements. Should we stick with our existing machine and juice up the RAM and add a faster hard drive? Any perspectives or advice welcome.
-
I have been playing around with a design pattern utilizing a table for storing layouts with everything else buoyed out. My question is in regards to performance. If I have 80+ tables with a few running several thousand records and growing Is this a bad idea. I would have less tables on the graph but more loading I presume of unnecessary tables on layouts that don't utilize them. Anyone, think I should avoid this strategy? Some layouts will have a thick patch of Tab Panels too, does related information get requested only when it is viewed? Not sure what goes on under the hood, some fead back would be greatly appreciated.
