Search Results

Search found 23576 results on 944 pages for 'case study'.

Page 130/944 | < Previous Page | 126 127 128 129 130 131 132 133 134 135 136 137  | Next Page >

  • Displaying device contacts with an indication that the contact is registered to the app

    - by Prasanna Aarthi
    We are developing a mobile app that needs to pick up device contacts, display them and indicate if the contact has already registered with this app. We have our DB in the server and the app fetches data using web services. What will be the best approach to implement the above scenario taking performance into consideration. Option 1: Every time user opens the app,fetch the contacts and send the list of email addresses to the server, check with the registered email ids and return the list of registered users in the contact list. In this approach whenever user opens the particular page, he needs to wait for few seconds to load data, but the contacts will be the latest from the device. Option 2: First time when the user opens the app, fetch contacts ,send the entire list of contacts and save it in the DB, retrieve list of registered users in the contacts then save this to local DB. From now on, data will be fetched from local DB and displayed. When a new user registers in the app, again check with records in central DB and send list of new users who are in your contacts that have registered to your app. This list will be added to local DB. and the process continues. In this case the new contacts added by user will not be updated in the app but retrieval and display of records would be quick. What would be the correct approach? In case there is a better way of doing this, please let me know.

    Read the article

  • SAP or Navision? Career Path

    - by codebased
    This could be tricky to ask; I may or may not ask this question here but I thought to give it a try. I've been in Software Industry since 2002 and now it has been a time that I'm at Senior level where I normally code, lead and define the architect; giving technical solutions to the management is one of my asset that I've earned during my services. Now it is the time to define the road map for the future, $$$. I am not in favor of Project Management roles. I've been thinking of going through the ERP and my current company does provide me an option to go for Navision/ Microsoft Dynamics. They are currently on 4.0 but they are planning to move for 2009 and also to build one of their own plug-in. Indeed the option is good because Microsoft is trying to accomplish the market for Dynamics products. However, they have less success in Australia. Now, Another option is with SAP where person can go with 200 K $ a year. Where as I'd doubt that if the same kind of growth, financial, is available for Microsoft geek. What is your opinion on Navision or SAP? If I try to completely move to SAP it could be bit challenging as market will consider me a fresher. However the return is quite good. Where in case of Microsoft, I think technology changes so fast that there is a less chance to grow in, within, the same experience; in other word, if any new framework comes in .net then market look for that person who knows this new framework and not .net But in case of SAP, where the base remain same and chances are to grab more money from the market. What would you do if you were me? In stackoverflow - Navision questions are 20+ where in SAP 200+///?? :-)

    Read the article

  • Electricity in VAIO laptop ports while plugged in

    - by SRbH
    I have VAIO E series (SVE) laptop. While charger plugged in, touching VGA/HDMI/USB port metal casing gives me burning sensation. So, i checked it with voltmeter and i found that while plugged in there is presences of DC +1.0 V. Same case with my friend's VAIO E Series (VPCEH). Service center says this is normal. But i do not believe them as it is not the case with Dell Inspiron. Please let me know if you are facing same problem with VAIO and is it hazardous to devices which i connect to these ports (like HDD, LED, etc)?

    Read the article

  • Creating a layer of abstraction over the ORM layer

    - by Daok
    I believe that if you have your repositories use an ORM that it's already enough abstracted from the database. However, where I am working now, someone believe that we should have a layer that abstract the ORM in case that we would like to change the ORM later. Is it really necessary or it's simply a lot of over head to create a layer that will work on many ORM? Edit Just to give more detail: We have POCO class and Entity Class that are mapped with AutoMapper. Entity class are used by the Repository layer. The repository layer then use the additional layer of abstraction to communicate with Entity Framework. The business layer has in no way a direct access to Entity Framework. Even without the additional layer of abstraction over the ORM, this one need to use the service layer that user the repository layer. In both case, the business layer is totally separated from the ORM. The main argument is to be able to change ORM in the future. Since it's really localized inside the Repository layer, to me, it's already well separated and I do not see why an additional layer of abstraction is required to have a "quality" code.

    Read the article

  • Command does not execute in crontab while command itself works just fine

    - by fuzzybee
    I have this script from Colin Johnson on Github - https://github.com/colinbjohnson/aws-missing-tools/tree/master/ec2-automate-backup It seems great. I have modified it to send email to myself every time an EBS snapshot is created or deleted. The following works like a charm ec2-automate-backup.sh -v "vol-myvolumeid" -k 3 However, it does not execute at all as part of my crontab (I didn't receive any emails) #some command that got commented out */5 * * * * ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3; * * * * * date /root/logs/crontab.log; */5 * * * * date /root/logs/crontab2.log Please note that the 2nd and 3rd execute just fines as I can see the date and time in log files. What could I have missed here? The full ec2-automate-backup.sh is as follows: #!/bin/bash - # Author: Colin Johnson / [email protected] # Date: 2012-09-24 # Version 0.1 # License Type: GNU GENERAL PUBLIC LICENSE, Version 3 # #confirms that executables required for succesful script execution are available prerequisite_check() { for prerequisite in basename ec2-create-snapshot ec2-create-tags ec2-describe-snapshots ec2-delete-snapshot date do #use of "hash" chosen as it is a shell builtin and will add programs to hash table, possibly speeding execution. Use of type also considered - open to suggestions. hash $prerequisite &> /dev/null if [[ $? == 1 ]] #has exits with exit status of 70, executable was not found then echo "In order to use `basename $0`, the executable \"$prerequisite\" must be installed." 1>&2 | mailx -s "Error happened 0" [email protected] ; exit 70 fi done } #get_EBS_List gets a list of available EBS instances depending upon the selection_method of EBS selection that is provided by user input get_EBS_List() { case $selection_method in volumeid) if [[ -z $volumeid ]] then echo "The selection method \"volumeid\" (which is $app_name's default selection_method of operation or requested by using the -s volumeid parameter) requires a volumeid (-v volumeid) for operation. Correct usage is as follows: \"-v vol-6d6a0527\",\"-s volumeid -v vol-6d6a0527\" or \"-v \"vol-6d6a0527 vol-636a0112\"\" if multiple volumes are to be selected." 1>&2 | mailx -s "Error happened 1" [email protected] ; exit 64 fi ebs_selection_string="$volumeid" ;; tag) if [[ -z $tag ]] then echo "The selected selection_method \"tag\" (-s tag) requires a valid tag (-t key=value) for operation. Correct usage is as follows: \"-s tag -t backup=true\" or \"-s tag -t Name=my_tag.\"" 1>&2 | mailx -s "Error happened 2" [email protected] ; exit 64 fi ebs_selection_string="--filter tag:$tag" ;; *) echo "If you specify a selection_method (-s selection_method) for selecting EBS volumes you must select either \"volumeid\" (-s volumeid) or \"tag\" (-s tag)." 1>&2 | mailx -s "Error happened 3" [email protected] ; exit 64 ;; esac #creates a list of all ebs volumes that match the selection string from above ebs_backup_list_complete=`ec2-describe-volumes --show-empty-fields --region $region $ebs_selection_string 2>&1` #takes the output of the previous command ebs_backup_list_result=`echo $?` if [[ $ebs_backup_list_result -gt 0 ]] then echo -e "An error occured when running ec2-describe-volumes. The error returned is below:\n$ebs_backup_list_complete" 1>&2 | mailx -s "Error happened 4" [email protected] ; exit 70 fi ebs_backup_list=`echo "$ebs_backup_list_complete" | grep ^VOLUME | cut -f 2` #code to right will output list of EBS volumes to be backed up: echo -e "Now outputting ebs_backup_list:\n$ebs_backup_list" } create_EBS_Snapshot_Tags() { #snapshot tags holds all tags that need to be applied to a given snapshot - by aggregating tags we ensure that ec2-create-tags is called only onece snapshot_tags="" #if $name_tag_create is true then append ec2ab_${ebs_selected}_$date_current to the variable $snapshot_tags if $name_tag_create then ec2_snapshot_resource_id=`echo "$ec2_create_snapshot_result" | cut -f 2` snapshot_tags="$snapshot_tags --tag Name=ec2ab_${ebs_selected}_$date_current" fi #if $purge_after_days is true, then append $purge_after_date to the variable $snapshot_tags if [[ -n $purge_after_days ]] then snapshot_tags="$snapshot_tags --tag PurgeAfter=$purge_after_date --tag PurgeAllow=true" fi #if $snapshot_tags is not zero length then set the tag on the snapshot using ec2-create-tags if [[ -n $snapshot_tags ]] then echo "Tagging Snapshot $ec2_snapshot_resource_id with the following Tags:" ec2-create-tags $ec2_snapshot_resource_id --region $region $snapshot_tags #echo "Snapshot tags successfully created" | mailx -s "Snapshot tags successfully created" [email protected] fi } date_command_get() { #finds full path to date binary date_binary_full_path=`which date` #command below is used to determine if date binary is gnu, macosx or other date_binary_file_result=`file -b $date_binary_full_path` case $date_binary_file_result in "Mach-O 64-bit executable x86_64") date_binary="macosx" ;; "ELF 64-bit LSB executable, x86-64, version 1 (SYSV)"*) date_binary="gnu" ;; *) date_binary="unknown" ;; esac #based on the installed date binary the case statement below will determine the method to use to determine "purge_after_days" in the future case $date_binary in gnu) date_command="date -d +${purge_after_days}days -u +%Y-%m-%d" ;; macosx) date_command="date -v+${purge_after_days}d -u +%Y-%m-%d" ;; unknown) date_command="date -d +${purge_after_days}days -u +%Y-%m-%d" ;; *) date_command="date -d +${purge_after_days}days -u +%Y-%m-%d" ;; esac } purge_EBS_Snapshots() { #snapshot_tag_list is a string that contains all snapshots with either the key PurgeAllow or PurgeAfter set snapshot_tag_list=`ec2-describe-tags --show-empty-fields --region $region --filter resource-type=snapshot --filter key=PurgeAllow,PurgeAfter` #snapshot_purge_allowed is a list of all snapshot_ids with PurgeAllow=true snapshot_purge_allowed=`echo "$snapshot_tag_list" | grep .*PurgeAllow'\t'true | cut -f 3` for snapshot_id_evaluated in $snapshot_purge_allowed do #gets the "PurgeAfter" date which is in UTC with YYYY-MM-DD format (or %Y-%m-%d) purge_after_date=`echo "$snapshot_tag_list" | grep .*$snapshot_id_evaluated'\t'PurgeAfter.* | cut -f 5` #if purge_after_date is not set then we have a problem. Need to alter user. if [[ -z $purge_after_date ]] #Alerts user to the fact that a Snapshot was found with PurgeAllow=true but with no PurgeAfter date. then echo "A Snapshot with the Snapshot ID $snapshot_id_evaluated has the tag \"PurgeAllow=true\" but does not have a \"PurgeAfter=YYYY-MM-DD\" date. $app_name is unable to determine if $snapshot_id_evaluated should be purged." 1>&2 | mailx -s "Error happened 5" [email protected] else #convert both the date_current and purge_after_date into epoch time to allow for comparison date_current_epoch=`date -j -f "%Y-%m-%d" "$date_current" "+%s"` purge_after_date_epoch=`date -j -f "%Y-%m-%d" "$purge_after_date" "+%s"` #perform compparison - if $purge_after_date_epoch is a lower number than $date_current_epoch than the PurgeAfter date is earlier than the current date - and the snapshot can be safely removed if [[ $purge_after_date_epoch < $date_current_epoch ]] then echo "The snapshot \"$snapshot_id_evaluated\" with the Purge After date of $purge_after_date will be deleted." ec2-delete-snapshot --region $region $snapshot_id_evaluated echo "Old snapshots successfully deleted for $volumeid" | mailx -s "Old snapshots successfully deleted for $volumeid" [email protected] fi fi done } #calls prerequisitecheck function to ensure that all executables required for script execution are available prerequisite_check app_name=`basename $0` #sets defaults selection_method="volumeid" region="ap-southeast-1" #date_binary allows a user to set the "date" binary that is installed on their system and, therefore, the options that will be given to the date binary to perform date calculations date_binary="" #sets the "Name" tag set for a snapshot to false - using "Name" requires that ec2-create-tags be called in addition to ec2-create-snapshot name_tag_create=false #sets the Purge Snapshot feature to false - this feature will eventually allow the removal of snapshots that have a "PurgeAfter" tag that is earlier than current date purge_snapshots=false #handles options processing while getopts :s:r:v:t:k:pn opt do case $opt in s) selection_method="$OPTARG";; r) region="$OPTARG";; v) volumeid="$OPTARG";; t) tag="$OPTARG";; k) purge_after_days="$OPTARG";; n) name_tag_create=true;; p) purge_snapshots=true;; *) echo "Error with Options Input. Cause of failure is most likely that an unsupported parameter was passed or a parameter was passed without a corresponding option." 1>&2 ; exit 64;; esac done #sets date variable date_current=`date -u +%Y-%m-%d` #sets the PurgeAfter tag to the number of days that a snapshot should be retained if [[ -n $purge_after_days ]] then #if the date_binary is not set, call the date_command_get function if [[ -z $date_binary ]] then date_command_get fi purge_after_date=`$date_command` echo "Snapshots taken by $app_name will be eligible for purging after the following date: $purge_after_date." fi #get_EBS_List gets a list of EBS instances for which a snapshot is desired. The list of EBS instances depends upon the selection_method that is provided by user input get_EBS_List #the loop below is called once for each volume in $ebs_backup_list - the currently selected EBS volume is passed in as "ebs_selected" for ebs_selected in $ebs_backup_list do ec2_snapshot_description="ec2ab_${ebs_selected}_$date_current" ec2_create_snapshot_result=`ec2-create-snapshot --region $region -d $ec2_snapshot_description $ebs_selected 2>&1` if [[ $? != 0 ]] then echo -e "An error occured when running ec2-create-snapshot. The error returned is below:\n$ec2_create_snapshot_result" 1>&2 ; exit 70 else ec2_snapshot_resource_id=`echo "$ec2_create_snapshot_result" | cut -f 2` echo "Snapshots successfully created for volume $volumeid" | mailx -s "Snapshots successfully created for $volumeid" [email protected] fi create_EBS_Snapshot_Tags done #if purge_snapshots is true, then run purge_EBS_Snapshots function if $purge_snapshots then echo "Snapshot Purging is Starting Now." purge_EBS_Snapshots fi cron log Oct 23 10:24:01 ip-10-130-153-227 CROND[28214]: (root) CMD (root (ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3;)) Oct 23 10:24:01 ip-10-130-153-227 CROND[28215]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:25:01 ip-10-130-153-227 CROND[28228]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:25:01 ip-10-130-153-227 CROND[28229]: (root) CMD (date >> /root/logs/crontab2.log) Oct 23 10:26:01 ip-10-130-153-227 CROND[28239]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:27:01 ip-10-130-153-227 CROND[28247]: (root) CMD (root (ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3;)) Oct 23 10:27:01 ip-10-130-153-227 CROND[28248]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:28:01 ip-10-130-153-227 CROND[28263]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:29:01 ip-10-130-153-227 CROND[28275]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:30:01 ip-10-130-153-227 CROND[28292]: (root) CMD (root (ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3;)) Oct 23 10:30:01 ip-10-130-153-227 CROND[28293]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:30:01 ip-10-130-153-227 CROND[28294]: (root) CMD (date >> /root/logs/crontab2.log) Oct 23 10:31:01 ip-10-130-153-227 CROND[28312]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:32:01 ip-10-130-153-227 CROND[28319]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:33:01 ip-10-130-153-227 CROND[28325]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:33:01 ip-10-130-153-227 CROND[28324]: (root) CMD (root (ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3;)) Oct 23 10:34:01 ip-10-130-153-227 CROND[28345]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:35:01 ip-10-130-153-227 CROND[28362]: (root) CMD (date >> /root/logs/crontab.log;) Oct 23 10:35:01 ip-10-130-153-227 CROND[28363]: (root) CMD (date >> /root/logs/crontab2.log) Mails to root From [email protected] Tue Oct 23 06:00:01 2012 Return-Path: <[email protected]> Date: Tue, 23 Oct 2012 06:00:01 GMT From: [email protected] (Cron Daemon) To: [email protected] Subject: Cron <root@ip-10-130-153-227> root ec2-automate-backup.sh -v "vol-fb2fbcdf" -k 3 Content-Type: text/plain; charset=UTF-8 Auto-Submitted: auto-generated X-Cron-Env: <SHELL=/bin/sh> X-Cron-Env: <HOME=/root> X-Cron-Env: <PATH=/usr/bin:/bin> X-Cron-Env: <LOGNAME=root> X-Cron-Env: <USER=root> Status: R /bin/sh: root: command not found

    Read the article

  • SOA Community Newsletter November 2012

    - by JuergenKress
    Dear SOA partner community member Too many different product from Oracle, no idea how do they fit together? Get a copy of the Oracle catalog, an excellent overview of the Oracle middleware portfolio. BPM is a key solution to this portfolio. To position BPM to your customers you can find many use case ideas in the paper BPM 11g Patterns and industry specific value propositions for Financial Services & Insurance & Retail. Many more Process Accelerators (11.1.1.6.2) have become available. It is an excellent demo and starting point for BPM projects. Our SOA Suite team published the most important OOW presentation at the OTN website. The Oracle SOA proactive support team is running a series of blog posts about SOA and JMS Introductory. To become an expert in SOA, Bob highlighted the latest list of SOA books. For OSB projects we recommend the EAIESB OSB poster. Thanks to all the experts who contributed and shared their SOA & BPM knowledge this month again. Please feel free to send us the link to your blog post via twitter @soacommunity: Undeploy multiple SOA composites with WLST or ANT by Danilo Schmiedel Fault Handling Slides and Q&A by Vennester Installing Oracle Event Processing 11g by Antoney Reynolds Expanding the Oracle Enterprise Repository with functional documentation by Marc Kuijpers Build Mobile App for E-Business Suite Using SOA Suite and ADF Mobile By Michelle Kimihira A brief note for customers running SOA Suite on AIX platforms By Christian ACM - Adaptive Case Management by Peter Paul BPM 11g - Dynamic Task Assignment with Multi-level Organization Units By Mark Foster Oracle Real User Experience Insight: Oracle's Approach to User Experience Hope to see you at the Middleware Day at UK Oracle User Group Conference 2012 in Birmingham. Jürgen Kress Oracle SOA & BPM Partner Adoption EMEA To read the newsletter please visit http://tinyurl.com/soanewsNovember2012 (OPN Account required) To become a member of the SOA Partner Community please register at http://www.oracle.com/goto/emea/soa (OPN account required) If you need support with your account please contact the Oracle Partner Business Center. Blog Twitter LinkedIn Mix Forum Technorati Tags: SOA Community newsletter,SOA Community,Oracle SOA,Oracle BPM,BPM Community,OPN,Jürgen Kress

    Read the article

  • SOA Community Newsletter November 2012

    - by JuergenKress
    Dear SOA partner community member Too many different product from Oracle, no idea how do they fit together? Get a copy of the Oracle catalog, an excellent overview of the Oracle middleware portfolio. BPM is a key solution to this portfolio. To position BPM to your customers you can find many use case ideas in the paper BPM 11g Patterns and industry specific value propositions for Financial Services & Insurance & Retail. Many more Process Accelerators (11.1.1.6.2) have become available. It is an excellent demo and starting point for BPM projects. Our SOA Suite team published the most important OOW presentation at the OTN website. The Oracle SOA proactive support team is running a series of blog posts about SOA and JMS Introductory. To become an expert in SOA, Bob highlighted the latest list of SOA books. For OSB projects we recommend the EAIESB OSB poster. Thanks to all the experts who contributed and shared their SOA & BPM knowledge this month again. Please feel free to send us the link to your blog post via twitter @soacommunity: Undeploy multiple SOA composites with WLST or ANT by Danilo Schmiedel Fault Handling Slides and Q&A by Vennester Installing Oracle Event Processing 11g by Antoney Reynolds Expanding the Oracle Enterprise Repository with functional documentation by Marc Kuijpers Build Mobile App for E-Business Suite Using SOA Suite and ADF Mobile By Michelle Kimihira A brief note for customers running SOA Suite on AIX platforms By Christian ACM - Adaptive Case Management by Peter Paul BPM 11g - Dynamic Task Assignment with Multi-level Organization Units By Mark Foster Oracle Real User Experience Insight: Oracle's Approach to User Experience Hope to see you at the Middleware Day at UK Oracle User Group Conference 2012 in Birmingham. Jürgen Kress Oracle SOA & BPM Partner Adoption EMEA To read the newsletter please visit http://tinyurl.com/soanewsNovember2012 (OPN Account required) To become a member of the SOA Partner Community please register at http://www.oracle.com/goto/emea/soa (OPN account required) If you need support with your account please contact the Oracle Partner Business Center. Blog Twitter LinkedIn Mix Forum Technorati Tags: SOA Community newsletter,SOA Community,Oracle SOA,Oracle BPM,BPM Community,OPN,Jürgen Kress

    Read the article

  • SQL SERVER – Storing Variable Values in Temporary Array or Temporary List

    - by pinaldave
    SQL Server does not support arrays or a dynamic length storage mechanism like list. Absolutely there are some clever workarounds and few extra-ordinary solutions but everybody can;t come up with such solution. Additionally, sometime the requirements are very simple that doing extraordinary coding is not required. Here is the simple case. Let us say here are the values: a, 10, 20, c, 30, d. Now the requirement is to store them in a array or list. It is very easy to do the same in C# or C. However, there is no quick way to do the same in SQL Server. Every single time when I get such requirement, I create a table variable and store the values in the table variables. Here is the example: For SQL Server 2012: DECLARE @ListofIDs TABLE(IDs VARCHAR(100)); INSERT INTO @ListofIDs VALUES('a'),('10'),('20'),('c'),('30'),('d'); SELECT IDs FROM @ListofIDs; GO When executed above script it will give following resultset. Above script will work in SQL Server 2012 only for SQL Server 2008 and earlier version run following code. DECLARE @ListofIDs TABLE(IDs VARCHAR(100), ID INT IDENTITY(1,1)); INSERT INTO @ListofIDs SELECT 'a' UNION ALL SELECT '10' UNION ALL SELECT '20' UNION ALL SELECT 'c' UNION ALL SELECT '30' UNION ALL SELECT 'd'; SELECT IDs FROM @ListofIDs; GO Now in this case, I have to convert numbers to varchars because I have to store mix datatypes in a single column. Additionally, this quick solution does not give any features of arrays (like inserting values in between as well accessing values using array index). Well, do you ever have to store temporary multiple values in SQL Server – if the count of values are dynamic and datatype is not specified early how will you about storing values which can be used later in the programming. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: PostADay, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology

    Read the article

  • How to visualize timer functionality in sequence diagram?

    - by truthseeker
    I am developing software for communication with external device through serial port. To better understand the new functionality I am trying to display it in sequence diagram. Flow of events is as follows. I send to the device command to reset it. This is asynchronous operation so there is some delay between request and response (typically 100 ms). There can be case when the answer never comes (for example device is not connected to the specified port or is currently turned off). For this purpose I create a timer with period twice the maximum answer time. In my case it is 2 * 125 ms = 250 ms. If the answer comes in predefined time interval, I destroy already running timer. If the answer doesnt come in predefined interval, timer initiates some action. After this action we can destroy it. How to effectively model this situation in sequence diagram? Addendum 1: Based on advices made by scarfridge i drew following UML diagram. Comment by Ozair is also helpful for simplifying the diagram even more.

    Read the article

  • Presenting the Cloud in a Different Way

    - by BuckWoody
    I had the honor of presenting the Developers at the Portland PASS chapter, and decided to go a different way than just using PowerPoint Slides…. (click on any picture to enlarge) The point is that when you need to get a point across, it’s OK to change tactics to make sure the information sticks. In this case, I decided to make the audience the PowerPoint. I used a few props to show the various paradigms we use to describe what the industry uses for the word “cloud” First, we talked about Infrastructure as a Service. I picked a gentleman who didn’t quite fit the hard hat and safety vest I picked out for him. Notice our “user” as she accesses our “Server” (complete with tray and glass) which has been virtualized.    Software as a service comes next. In this case, the user and potentially even customers just access software (represented here as a Windows ME box…) remotely – everything is virtualized. Finally, Platform as a Service – Yup, Platform shoes as a necklace, and a tie-dye shirt to represent the 70’s – a decade when mainframes used stateless programming as well. Notice also the components of Windows Azure – Compute (Keyboard) Application Fabric (Toy Bus) and Storage (Bucket).   And at the end of the day, it’s all about serving those customers…

    Read the article

  • How should I deal with user agent parsing in logs?

    - by Mr. Jefferson
    My web app project includes logging functionality so we can see where visitors are coming from (referrer URL), what the popular user agents are, what pages are most popular, etc. The log is stored in SQL Server, and when I query the user agents I use a large (almost 100 lines) and growing CASE statement to separate the user agents using string matching (i.e. if the user agent contains the string "Firefox/9" then it's Firefox 9). Is there a better way to do this so I don't have to continually add to that CASE statement to deal with new browser releases? Also, how should I deal with less common, weird/unknown user agents? I've seen the following in the logs and been unable to find good information online about what they are: WordPress/3.3.1; http://www.facecolony.org Mozilla/4.0 ( http://www.hairirons.org redips; <a href=http://hairirons.org/>chi hair iron</a>) I'd guess they're bots/crawlers, but the sites they point to don't appear to reference web crawlers (or even be available sometimes). I've seen other user agents aren't familiar to me, but I know they're bots because they include "bot" or "spider" or something similar in them.

    Read the article

  • Using orientation to calculate position on Windows Phone 7

    - by Lavinski
    I'm using the motion API and I'm trying to figure out a control scheme for the game I'm currently developing. What I'm trying to achive is for a orienation of the device to correlate directly to a position. Such that tilting the phone forward and to the left represents the top left position and back to the right would be the bottom right position. Photos to make it clearer (the red dot would be the calculated position). Forward and Left Back and Right Now for the tricky bit. I also have to make sure that the values take into account left landscape and right landscape device orientations (portrait is the default so no calculations would be needed for it). Has anyone done anything like this? Notes: I've tried using the yaw, pitch, roll and Quaternion readings. Sample: // Get device facing vector public static Vector3 GetState() { lock (lockable) { var down = Vector3.Forward; var direction = Vector3.Transform(down, state); switch (Orientation) { case Orientation.LandscapeLeft: return Vector3.TransformNormal(direction, Matrix.CreateRotationZ(-rightAngle)); case Orientation.LandscapeRight: return Vector3.TransformNormal(direction, Matrix.CreateRotationZ(rightAngle)); } return direction; } }

    Read the article

  • Creating country specific twitter/facebook accounts

    - by user359650
    I see many companies that have an international presence trying to localize their social media presence by creating country or language specific accounts. However some seemed to have done so without following a consistent pattern, one example being the World Wildlife Fund when you look at their Twitter accounts: World_Wildlife : verified account with 200K followers WWF : main account with 800K followers www_uk : lower case with underscore between WWF and country indicator WWFCanada : upper case with country indicator attached to WWF ... I am planning to build a website which hopefully will grow global and would like to avoid this sort of inconsistencies. Also, I was comparing what Twitter and Facebook allow in their username and found out that they don't allow the same characters to be used (e.g. for instance that the former doesn't allow . whereas the latter does) making difficult to ensure consistency across social networks. Hence my questions: Are there known naming schemes for creating localized Twitter and Facebook accounts while maintaining a certain consistency between them (best effort)? Are there any researches out there that have proven whether some schemes were better than others in terms of readability and/or SEO?

    Read the article

  • Drbd Primary/Primary + iSCSI: accessing to different files avoids split brain?

    - by Eddie C.
    I have a question / curiosity about split-brain on a Drbd Primary/Primary configuration. Supposing two nodes (hosts), host1 and host2 configured with Drbd Primary/Primary and two different shares (NFS, CIFS o iSCSI) of a replicated area (saying /drbd) /drbd/file1.data /drbd/file2.data If a pool of client would access only by host1 share reading and wrinting only file1.data and another pool only by host2 share to file2.data, this scenario should avoid split brain situation in case of one node failure or it's just a conjecture? The final purpose is load balance between the two nodes in normal condition and collapsing to one node only in case of failure. Thank you! Eddie

    Read the article

  • ArchBeat Link-o-Rama for 2012-09-20

    - by Bob Rhubart
    Attend OTN Architect Day – by Architects, for Architects – October 25 You won't need 3D glasses to take in these live presentations (8 sessions, two tracks) on Cloud computing, SOA, and engineered systems. And the ticket price is: Zero. Nothing. Absolutely free. Register now for Oracle Technology Network Architect Day in Los Angeles. Thursday October 25, 2012, 8:00am – 5:00pm Sofitel Los Angeles8555 Beverly BoulevardLos Angeles, CA 90048 Loving VirtualBox 4.2… | The ORACLE-BASE Blog Is it wrong for a man to love a technology? Oracle ACE Director Tim Hall has several very good reasons for his feelings… Running RichFaces on WebLogic 12c | Markus Eisele "With all the JMS magic and the different provider checks in the showcase this has become some kind of a challenge to simply build and deploy it," says Oracle ACE Director Markus Eisele. His detailed post will help you to meet that challenge. Oracle ADF Coverage at OOW | Frank Nimphius Frank Nimphius shares a comprehensive and well-organized list of Oracle ADF sessions and activities scheduled for Oracle OpenWorld in San Francisco. OIM 11g R2 Catalog Customization Example | Daniel Gralewski Oracle Fusion Middleware A-Team member Daniel Gralewski's post shows "how OIM catalog can be customized by using OIM UI capabilities such as managed beans and EL expressions. The post first describes the use case and the solution to address the use case; then it describes the solution details as well as provides links to the artifacts." New Book: Oracle BPM Suite 11g: Advanced BPMN Topics | Mark Nelson Redstack blogger Mark Nelson shares an overview of Oracle BPM Suite 11g: Advanced BPMN Topics, the new book he co-authored with Tanya Williams. Nelson describes the book as "a concise presentation of both theory and practical examples of the areas of BPMN where we have encountered the most widespread confusion and misunderstanding." Thought for the Day "I strive for an architecture from which nothing can be taken away." — Helmut Jahn Source: Brainy Quote

    Read the article

  • E.T. Phone "Home" - Hey I've discovered a leak..!

    - by Martin Deh
    Being a member of the WebCenter ATEAM, we are often asked to performance tune a WebCenter custom portal application or a WebCenter Spaces deployment.  Most of the time, the process is pretty much the same.  For example, we often use tools like httpWatch and FireBug to monitor the application, and then perform load tests using JMeter or Selenium.  In addition, there are the fine tuning of the different performance based tuning parameters that are outlined in the documentation and by blogs that have been written by my fellow ATEAMers (click on the "performance" tag in this ATEAM blog).  While performing the load test where the outcome produces a significant reduction in the systems resources (memory), one of the causes that plays a role in memory "leakage" is due to the implementation of the navigation menu UI.  OOTB in both JDeveloper and WebCenter Spaces, there are sample (page) templates that include a "default" navigation menu.  In WebCenter Spaces, this is through the SpacesNavigationModel taskflow region, and in a custom portal (i.e. pageTemplate_globe.jspx) the menu UI is contructed using standard ADF components.  These sample menu UI's basically enable the underlying navigation model to visualize itself to some extent.  However, due to certain limitations of these sample menu implementations (i.e. deeper sub-level of navigations items, look-n-feel, .etc), many customers have developed their own custom navigation menus using a combination of HTML, CSS and JQuery.  While this is supported somewhat by the framework, it is important to know what are some of the best practices in ensuring that the navigation menu does not leak.  In addition, in this blog I will point out a leak (BUG) that is in the sample templates.  OK, E.T. the suspence is killing me, what is this leak? Note: for those who don't know, info on E.T. can be found here In both of the included templates, the example given for handling the navigation back to the "Home" page, will essentially provide a nice little memory leak every time the link is clicked. Let's take a look a simple example, which uses the default template in Spaces. The outlined section below is the "link", which is used to enable a user to navigation back quickly to the Group Space Home page. When you (mouse) hover over the link, the browser displays the target URL. From looking initially at the proposed URL, this is the intended destination.  Note: "home" in this case is the navigation model reference (id), that enables the display of the "pretty URL". Next, notice the current URL, which is displayed in the browser.  Remember, that PortalSiteHome = home.  The other highlighted item adf.ctrl-state, is very important to the framework.  This item is basically a persistent query parameter, which is used by the (ADF) framework to managing the current session and page instance.  Without this parameter present, among other things, the browser back-button navigation will fail.  In this example, the value for this parameter is currently 95K25i7dd_4.  Next, through the navigation menu item, I will click on the Page2 link. Inspecting the URL again, I can see that it reports that indeed the navigation is successful and the adf.ctrl-state is also in the URL.  For those that are wondering why the URL displays Page3.jspx, instead of Page2.jspx. Basically the (file) naming convention for pages created ar runtime in Spaces start at Page1, and then increment as you create additional pages.  The name of the actual link (i.e. Page2) is the page "title" attribute.  So the moral of the story is, unlike design time created pages, run time created pages the name of the file will 99% never match the name that appears in the link. Next, is to click on the quick link for navigating back to the Home page. Quick investigation yields that the navigation was indeed successful.  In the browser's URL there is a home (pretty URL) reference, and there is also a reference to the adf.ctrl-state parameter.  So what's the issue?  Can you remember what the value was for the adf.ctrl-state?  The current value is 3D95k25i7dd_149.  However, the previous value was 95k25i7dd_4.  Here is what happened.  Remember when (mouse) hovering over the link produced the following target URL: http://localhost:8888/webcenter/spaces/NavigationTest/home This is great for the browser as this URL will navigate to the intended targer.  However, what is missing is the adf.ctrl-state parameter.  Since this parameter was not present upon navigation "within" the framework, the ADF framework produced another adf.ctrl-state (object).  The previous adf.ctrl-state basically is orphaned while continuing to be alive in memory.  Note: the auto-creation of the adf.ctrl state does happen initially when you invoke the Spaces application  for the first time.  The following is the line of code which produced the issue: <af:goLink destination="#{boilerBean.globalLogoURIInSpace} ... Here the boilerBean is responsible for returning the "string" url, which in this case is /spaces/NavigationTest/home. Unfortunately, again what is missing is adf.ctrl-state. Note: there are more than one instance of the goLinks in the sample templates. So E.T. how can I correct this? There are 2 simple fixes.  For the goLink's destination, use the navigation model to return the actually "node" value, then use the goLinkPrettyUrl method to add the current adf.ctrl-state: <af:goLink destination="#{navigationContext.defaultNavigationModel.node['home'].goLinkPrettyUrl}"} ... />  Note: the node value is the [navigation model id]  Using a goLink does solve the main issue.  However, since the link basically does a redirect, some browsers like IE will produce a somewhat significant "flash".  In a Spaces application, this may be an annoyance to the users.  Another way to solve the leakage problem, and also remove the flash between navigations is to use a af:commandLink.  For example, here is the code example for this scenario: <af:commandLink id="pt_cl2asf" actionListener="#{navigationContext.processAction}" action="pprnav">    <f:attribute name="node" value="#{navigationContext.defaultNavigationModel.node['home']}"/> </af:commandLink> Here, the navigation node to where home is located is delivered by way of the attribute to the commandLink.  The actual navigation is performed by the processAction, which is needing the "node" value. E.T. OK, you solved the OOTB sample BUG, what about my custom navigation code? I have seen many implementations of creating a navigation menu through custom code.  In addition, there are some blog sites that also give detailed examples.  The majority of these implementations are very similar.  The code usually involves using standard HTML tags (i.e. DIVS, UL, LI, .,etc) and either CSS or JavaScript (JQuery) to produce the flyout/drop-down effect.  The navigation links in these cases are standard <a href... > tags.  Although, this type of approach is not fully accepted by the ADF community, it does work.  The important thing to note here is that the <a> tag value must use the goLinkPrettyURL method of contructing the target URL.  For example: <a href="${contextRoot}${menu.goLinkPrettyUrl}"> The main reason why this type of approach is popular is that links that are created this way (also with using af:goLinks), the pages become crawlable by search engines.  CommandLinks are currently not search friendly.  However, in the case of a Spaces instance this may be acceptable.  So in this use-case, af:commandLinks, which would replace the <a>  (or goLink) tags. The example code given of the af:commandLink above is still valid. One last important item.  If you choose to use af:commandLinks, special attention must be given to the scenario in which java script has been used to produce the flyout effect in the custom menu UI.  In many cases that I have seen, the commandLink can only be invoked once, since there is a conflict between the custom java script with the ADF frameworks own scripting to control the view.  The recommendation here, would be to use a pure CSS approach to acheive the dropdown effects. One very important thing to note.  Due to another BUG, the WebCenter environement must be patched to BP3 (patch  p14076906).  Otherwise the leak is still present using the goLinkPrettyUrl method.  Thanks E.T.!  Now I can phone home and not worry about my application running out of resources due to my custom navigation! 

    Read the article

  • Positioning a sprite in XNA: Use ClientBounds or BackBuffer?

    - by Martin Andersson
    I'm reading a book called "Learning XNA 4.0" written by Aaron Reed. Throughout most of the chapters, whenever he calculates the position of a sprite to use in his call to SpriteBatch.Draw, he uses Window.ClientBounds.Width and Window.ClientBounds.Height. But then all of a sudden, on page 108, he uses PresentationParameters.BackBufferWidth and PresentationParameters.BackBufferHeight instead. I think I understand what the Back Buffer and the Client Bounds are and the difference between those two (or perhaps not?). But I'm mighty confused about when I should use one or the other when it comes to positioning sprites. The author uses for the most part Client Bounds both for checking whenever a moving sprite is of the screen and to find a spawn point for new sprites. However, he seems to make two exceptions from this pattern in his book. The first time is when he wants some animated sprites to "move in" and cross the screen from one side to another (page 108 as mentioned). The second and last time is when he positions a texture to work as a button in the lower right corner of a Windows Phone 7 screen (page 379). Anyone got an idea? I shall provide some context if it is of any help. Here's how he usually calls SpriteBatch.Draw (code example from where he positions a sprite in the middle of the screen [page 35]): spriteBatch.Draw(texture, new Vector2( (Window.ClientBounds.Width / 2) - (texture.Width / 2), (Window.ClientBounds.Height / 2) - (texture.Height / 2)), null, Color.White, 0, Vector2.Zero, 1, SpriteEffects.None, 0); And here is the first case of four possible in a switch statement that will set the position of soon to be spawned moving sprites, this position will later be used in the SpriteBatch.Draw call (page 108): // Randomly choose which side of the screen to place enemy, // then randomly create a position along that side of the screen // and randomly choose a speed for the enemy switch (((Game1)Game).rnd.Next(4)) { case 0: // LEFT to RIGHT position = new Vector2( -frameSize.X, ((Game1)Game).rnd.Next(0, Game.GraphicsDevice.PresentationParameters.BackBufferHeight - frameSize.Y)); speed = new Vector2(((Game1)Game).rnd.Next( enemyMinSpeed, enemyMaxSpeed), 0); break;

    Read the article

  • Oracle@info360: Advance Beyond Point Solutions To An Enterprise Content Strategy

    - by kellsey.ruppel(at)oracle.com
    The info360/AIIM conference is March 22-24 in Washington DC. We have a number of customer speakers this year talking on the theme of “Advance Beyond Point Solutions To An Enterprise Content Strategy.” These customers all started by addressing a particular use case, but then used the infrastructure they had created to quickly and cost effectively stand up solutions to new business problems.  Andy MacMillan, VP of Product Management at Oracle, will give a thought provoking opening keynote at 8:50 AM on Tuesday, March 22nd. He will be joined by Juan Jose Goldschtein, the CIO of the Organization of American States. The OAS has developed a human rights website that is the front end to a case management system for human rights violations. The implementation supports digital signatures on iPads, so their executives can approve workflows and keep cases moving forward while they are busy traveling and investigating abuses.Other customer speakers include:Tom Robinette, Director of Applications and IT Engineering, Dresser-RandRobin Crisp, Program Manager, FDAMonica Crocker, Corporate Records Manager, Land O’ LakesBrian Skapura, The American Institute of ArchitectsKathy Adams and Leslie Becker, The Nature ConservancyIrfan Motiwala, Sr. VP, Moody’s Investment ServicesMolly Wenzler, Director of Electronic Media, MeadWestvaco Other sessions include our Super Session that kicks off the Oracle Track @info360 on Wednesday. At 11:00 AM, Senior Director of Product Marketing, Howard Beader will present The Social Enterprise – Combining People, Processes and Content. This session will focus on how customers have brought social media, business process management, and content management together to supercharge their organizations. Oracle customers can arrange one-on-one meetings with Oracle executives and product experts, and attend the VIP customer appreciation event. Oracle will be joined by Oracle partners:FujitsuKesteTeamInformaticsKapowSena SystemsDTIYou can learn more about discounts for Oracle customers and register on our Oracle@info360 page.To see more about the customers and sessions that will be presented, you can look at the Oracle Track page on the AIIM/info360 website.Technorati Tags: oracle, AIIM, info360, content management, social enterprise

    Read the article

  • Accessing the JSESSIONID from JSF

    - by Frank Nimphius
    The following code attempts to access and print the user session ID from ADF Faces, using the session cookie that is automatically set by the server and the Http Session object itself. FacesContext fctx = FacesContext.getCurrentInstance(); ExternalContext ectx = fctx.getExternalContext(); HttpSession session = (HttpSession) ectx.getSession(false); String sessionId = session.getId(); System.out.println("Session Id = "+ sessionId); Cookie[] cookies = ((HttpServletRequest)ectx.getRequest()).getCookies(); //reset session string sessionId = null; if (cookies != null) { for (Cookie brezel : cookies) {     if (brezel.getName().equalsIgnoreCase("JSESSIONID")) {        sessionId = brezel.getValue();        break;      }   } } System.out.println("JSESSIONID cookie = "+sessionId); Though apparently both approaches to the same thing, they are different in the value they return and the condition under which they work. The getId method, for example returns a session value as shown below grLFTNzJhhnQTqVwxHMGl0WDZPGhZFl2m0JS5SyYVmZqvrfghFxy!-1834097692!1322120041091 Reading the cookie, returns a value like this grLFTNzJhhnQTqVwxHMGl0WDZPGhZFl2m0JS5SyYVmZqvrfghFxy!-1834097692 Though both seem to be identical, the difference is within "!1322120041091" added to the id when reading it directly from the Http Session object. Dependent on the use case the session Id is looked up for, the difference may not be important. Another difference however, is of importance. The cookie reading only works if the session Id is added as a cookie to the request, which is configurable for applications in the weblogic-application.xml file. If cookies are disabled, then the server adds the session ID to the request URL (actually it appends it to the end of the URI, so right after the view Id reference). In this case however no cookie is set so that the lookup returns empty. In both cases however, the getId variant works.

    Read the article

  • Open World Session - BPM, SOA and ADF Combined:Patterns learned from Fusion Applications

    - by mesriniv
    Blog by Meera Srinivasan (Oracle Product Management) Today afternoon (10/2/2012), Mohan Kamath, and I (Meera Srinivasan) delivered an Open World session on how Oracle Fusion Applications (the next generation business applications from Oracle), use Oracle BPM, Oracle SOA and Oracle ADF products. These adoption patterns can be applied in a generic manner to produce process-centric, user-centric, highly customizable and extensible next generation application. The session was well attended and we had lively discussions with the attendees during Q & A. We started with why as an application developer, you should look at BPM for creating a process-centric application and presented the following fusion adoption patterns Model driven agile development Customization and Extension Guided Process Interactions Personalization and Customization of End User Interfaces Approval Flows Fusion HCM, On Boarding Process - Activity Guide Interface was used as an example for the Guided Process Interactions adoption pattern and the Fusion CRM BPM Process Templates for Customization adoption pattern. In the Personalization and Customization of End User Interfaces section, we looked at how ADF is used within Oracle BPM and the various options available to customize end user interfaces. We also presented how Oracle Procurement does complex approvals using Rules and Approval Management Extensions. We hope you found the session useful, and please do try to attend Heidi’s session on dynamic case management: Case Management Patterns with Oracle Unified Business Process Management Suite. Marriott Marquis - Salon 7, Thu 11:15 AM - 12:15 PM

    Read the article

  • Why do my gvfs mounts not show up under ~/.gvfs?

    - by kynan
    From what I read, when mounting a network share via nautilus or gvfs-mount the mount point should be in ~/.gvfs. This seems not to be the case for me: I tried mounting both an FTP and SMB share via both nautilus and gvfs-mount under both Ubuntu Maverick and Natty and in none of the cases did I see any mount point under ~/.gvfs. I can access the shares just find in nautilus, but I want to have access via the command line, which is why I need a mount point in the file system. Edit: Debugging following James Henstridge's answer and enzotib's comment revealed that on my laptop gvfs-fuse-daemon is running and consequently gvfs mounts show up in ~/.gvfs, whereas on the 2 workstations where ~/.gvfs remained empty gvfs-fuse-daemon was not running. On all 3 machines there are other gvfs processes running: gvfsd, gvfs-afc-volume-monitor, ... On the laptop, mount | fgrep gvfs yields gvfs-fuse-daemon on /home/xxx/.gvfs type fuse.gvfs-fuse-daemon (rw,nosuid,nodev,user=xxx) That raises the questions: How are shares mounted without gvfs-fuse-daemon running? Is there no mount point created in that case and is every access to the share a gvfs library call? Which daemon is responsible? gvfsd? What's the role of gvfs-fuse-daemon? Does it only create a fuse mount point in ~/.gvfs?

    Read the article

  • Recommended total system backup solution

    - by bioShark
    I hope this question won't get closed immediately since it has a generic title. I already searched a bit around the answers here, but nothing satisfied me. I want a back-up solution that makes a total back-up, so that I can restore my Ubuntu in case of major failures, like HDD failing. As far as I can see, I have 2 choices: 1) Backing up with Deja Dup to an external disk. This is fine and I am already doing, but in case my HDD fails, and I make a new Ubuntu install on a new disk, will Deja Dup be able to restore all my setting and stuff from the backed up files? If it can, then what other files/folders should I add in Deja Dup to back-up (currently I have set only the recommended /home folder)? Is there a point in telling Deja Dup to back-up everything under "/" ? 2) A disk/partition cloning software. This would be something similar to Noton ghost. Is there such a software with nice GUI that you could recommend for Ubuntu? And even better, it would be nice if Ubuntu's liveCD could recognize such a clone at install step. I am using 11.10

    Read the article

  • OPN Exchange Keynote On-Demand

    - by kristin.jellison
    We hope everyone has had a chance to refresh and recharge after Oracle OpenWorld 2013. In case you didn’t have the opportunity to catch the full OPN Exchange keynote, we have it on demand for your viewing pleasure. A highlight reel is up on the OPN YouTube channel and on Oracle.com. You can also watch individual keynote segments, from Oracle Executives like Mark Hurd, John Fowler and Andy Bailey, highlighted below. So please, sit back, relax and enjoy the show! You know, in case your football team is on a bye this week. Mark Hurd, President, Oracle Executive Address John Fowler, Executive Vice President, Systems Hardware and Software Engineered to Work Together Joel Borellis, Group Vice President, Partner Enablement Technology, Middleware and Business Intelligence Chris Baker, Senior Vice President, Worldwide ISV,OEM and Java Sales Engineered Systems and Hardware Andy Bailey, Senior Vice President, Strategic Alliances Cloud, Fusion Applications and Customer Experience Thomas LaRocca, Senior Vice President, North America Sales Alliances and Channels Terri Hall, Group Vice President, North America Sales Alliances and Channels Oracle Partner Excellence Awards: North America Hugo Freytes, Senior Vice President, Latin America Alliances and Channels Oracle Partner Excellence Awards: Latin America Mark Lewis, Senior Vice President, APAC Alliances and Channels Hiroshi Watanabe, Senior Vice President, Japan Alliances and Channels Oracle Partner Excellence Awards: APAC and Japan David Callaghan, Senior Vice President, EMEA Alliances and Channels Oracle Partner Excellence Awards: EMEA Cheers! The OPN Communications Team

    Read the article

  • Parallel Classloading Revisited: Fully Concurrent Loading

    - by davidholmes
    Java 7 introduced support for parallel classloading. A description of that project and its goals can be found here: http://openjdk.java.net/groups/core-libs/ClassLoaderProposal.html The solution for parallel classloading was to add to each class loader a ConcurrentHashMap, referenced through a new field, parallelLockMap. This contains a mapping from class names to Objects to use as a classloading lock for that class name. This was then used in the following way: protected Class loadClass(String name, boolean resolve) throws ClassNotFoundException { synchronized (getClassLoadingLock(name)) { // First, check if the class has already been loaded Class c = findLoadedClass(name); if (c == null) { long t0 = System.nanoTime(); try { if (parent != null) { c = parent.loadClass(name, false); } else { c = findBootstrapClassOrNull(name); } } catch (ClassNotFoundException e) { // ClassNotFoundException thrown if class not found // from the non-null parent class loader } if (c == null) { // If still not found, then invoke findClass in order // to find the class. long t1 = System.nanoTime(); c = findClass(name); // this is the defining class loader; record the stats sun.misc.PerfCounter.getParentDelegationTime().addTime(t1 - t0); sun.misc.PerfCounter.getFindClassTime().addElapsedTimeFrom(t1); sun.misc.PerfCounter.getFindClasses().increment(); } } if (resolve) { resolveClass(c); } return c; } } Where getClassLoadingLock simply does: protected Object getClassLoadingLock(String className) { Object lock = this; if (parallelLockMap != null) { Object newLock = new Object(); lock = parallelLockMap.putIfAbsent(className, newLock); if (lock == null) { lock = newLock; } } return lock; } This approach is very inefficient in terms of the space used per map and the number of maps. First, there is a map per-classloader. As per the code above under normal delegation the current classloader creates and acquires a lock for the given class, checks if it is already loaded, then asks its parent to load it; the parent in turn creates another lock in its own map, checks if the class is already loaded and then delegates to its parent and so on till the boot loader is invoked for which there is no map and no lock. So even in the simplest of applications, you will have two maps (in the system and extensions loaders) for every class that has to be loaded transitively from the application's main class. If you knew before hand which loader would actually load the class the locking would only need to be performed in that loader. As it stands the locking is completely unnecessary for all classes loaded by the boot loader. Secondly, once loading has completed and findClass will return the class, the lock and the map entry is completely unnecessary. But as it stands, the lock objects and their associated entries are never removed from the map. It is worth understanding exactly what the locking is intended to achieve, as this will help us understand potential remedies to the above inefficiencies. Given this is the support for parallel classloading, the class loader itself is unlikely to need to guard against concurrent load attempts - and if that were not the case it is likely that the classloader would need a different means to protect itself rather than a lock per class. Ultimately when a class file is located and the class has to be loaded, defineClass is called which calls into the VM - the VM does not require any locking at the Java level and uses its own mutexes for guarding its internal data structures (such as the system dictionary). The classloader locking is primarily needed to address the following situation: if two threads attempt to load the same class, one will initiate the request through the appropriate loader and eventually cause defineClass to be invoked. Meanwhile the second attempt will block trying to acquire the lock. Once the class is loaded the first thread will release the lock, allowing the second to acquire it. The second thread then sees that the class has now been loaded and will return that class. Neither thread can tell which did the loading and they both continue successfully. Consider if no lock was acquired in the classloader. Both threads will eventually locate the file for the class, read in the bytecodes and call defineClass to actually load the class. In this case the first to call defineClass will succeed, while the second will encounter an exception due to an attempted redefinition of an existing class. It is solely for this error condition that the lock has to be used. (Note that parallel capable classloaders should not need to be doing old deadlock-avoidance tricks like doing a wait() on the lock object\!). There are a number of obvious things we can try to solve this problem and they basically take three forms: Remove the need for locking. This might be achieved by having a new version of defineClass which acts like defineClassIfNotPresent - simply returning an existing Class rather than triggering an exception. Increase the coarseness of locking to reduce the number of lock objects and/or maps. For example, using a single shared lockMap instead of a per-loader lockMap. Reduce the lifetime of lock objects so that entries are removed from the map when no longer needed (eg remove after loading, use weak references to the lock objects and cleanup the map periodically). There are pros and cons to each of these approaches. Unfortunately a significant "con" is that the API introduced in Java 7 to support parallel classloading has essentially mandated that these locks do in fact exist, and they are accessible to the application code (indirectly through the classloader if it exposes them - which a custom loader might do - and regardless they are accessible to custom classloaders). So while we can reason that we could do parallel classloading with no locking, we can not implement this without breaking the specification for parallel classloading that was put in place for Java 7. Similarly we might reason that we can remove a mapping (and the lock object) because the class is already loaded, but this would again violate the specification because it can be reasoned that the following assertion should hold true: Object lock1 = loader.getClassLoadingLock(name); loader.loadClass(name); Object lock2 = loader.getClassLoadingLock(name); assert lock1 == lock2; Without modifying the specification, or at least doing some creative wordsmithing on it, options 1 and 3 are precluded. Even then there are caveats, for example if findLoadedClass is not atomic with respect to defineClass, then you can have concurrent calls to findLoadedClass from different threads and that could be expensive (this is also an argument against moving findLoadedClass outside the locked region - it may speed up the common case where the class is already loaded, but the cost of re-executing after acquiring the lock could be prohibitive. Even option 2 might need some wordsmithing on the specification because the specification for getClassLoadingLock states "returns a dedicated object associated with the specified class name". The question is, what does "dedicated" mean here? Does it mean unique in the sense that the returned object is only associated with the given class in the current loader? Or can the object actually guard loading of multiple classes, possibly across different class loaders? So it seems that changing the specification will be inevitable if we wish to do something here. In which case lets go for something that more cleanly defines what we want to be doing: fully concurrent class-loading. Note: defineClassIfNotPresent is already implemented in the VM as find_or_define_class. It is only used if the AllowParallelDefineClass flag is set. This gives us an easy hook into existing VM mechanics. Proposal: Fully Concurrent ClassLoaders The proposal is that we expand on the notion of a parallel capable class loader and define a "fully concurrent parallel capable class loader" or fully concurrent loader, for short. A fully concurrent loader uses no synchronization in loadClass and the VM uses the "parallel define class" mechanism. For a fully concurrent loader getClassLoadingLock() can return null (or perhaps not - it doesn't matter as we won't use the result anyway). At present we have not made any changes to this method. All the parallel capable JDK classloaders become fully concurrent loaders. This doesn't require any code re-design as none of the mechanisms implemented rely on the per-name locking provided by the parallelLockMap. This seems to give us a path to remove all locking at the Java level during classloading, while retaining full compatibility with Java 7 parallel capable loaders. Fully concurrent loaders will still encounter the performance penalty associated with concurrent attempts to find and prepare a class's bytecode for definition by the VM. What this penalty is depends on the number of concurrent load attempts possible (a function of the number of threads and the application logic, and dependent on the number of processors), and the costs associated with finding and preparing the bytecodes. This obviously has to be measured across a range of applications. Preliminary webrevs: http://cr.openjdk.java.net/~dholmes/concurrent-loaders/webrev.hotspot/ http://cr.openjdk.java.net/~dholmes/concurrent-loaders/webrev.jdk/ Please direct all comments to the mailing list [email protected].

    Read the article

< Previous Page | 126 127 128 129 130 131 132 133 134 135 136 137  | Next Page >