Network.http.max-persistent-connections-per-server keeps reverting to user set 4

I didn't used to have this problem as far as I'm aware, but as of recently, I've noticed that if I try to download more than 4 of one thing at a time from the same site, it would fail. For instance, if I have up 5 web pages of streaming content, only 4 will load, the other 5th one will do nothing until I finish one of the others. Upon looking up some information on this, I saw that the network.http.max-persistent-connections-per-server is set at "user set = 4", and if I change that number to anything else, it corresponds. If I click to 'reset' the number, it goes to the default of 6 - which is still less than what it used to be. I used to not even have a limit on this as far as I'm aware, or if I did, it was well beyond 10 and that wasn't a product of me manipulating it.
What I'm asking is this:
1) How do I make this so it permanently switches to a higher number? Every time I reset it to 6 or I manually change the number to something higher, once I close it, those settings go back to 4.
2) Why is it reverting and not staying the way I change it?
3) What's the purpose of this and why is it just recently doing it when I haven't put any new add-ons or made any other refinements to Firefox that I'm aware of?
Any help would be appreciated

Two possible causes for it resetting to 4 that I can think of.
# You have an add-on handing that pref and that is resetting it to 4 when Firefox is restarted.
# That pref is set via user.js and that causes it to go to that value as Firefox starts.
http://support.mozilla.com/en-US/kb/troubleshooting+extensions+and+themes <br />
http://kb.mozillazine.org/User.js_file

Similar Messages

I can't cap the number of active downloads (via persistent-connection-to-server cap) AND still be able to load pages hosted on the same server, *because*...?

I'm aware that, by setting "network.http.max-persistent-connections-per-server" (in about:config) to 1, I'm effectively preventing myself from being able to download a single file from a server AND load a page hosted on the same server simultaneously.
My question is, why can't Firefox tell the difference between a file download and loading a website?
The reason I ask is that I feel I should be able to create a download queue and have my browser download one file at a time (to minimize the impact downloads have on available bandwidth and my own downstream, while allowing me to set up a long chain of sequential downloads), YET still be able to browse websites hosted on the same server.
I do not want to use an add-on or a separate download manager to solve this problem. I think that's stupid. Firefox should be able to tell the difference between a request to load an html file and a request to download any other type of file.
After all, I can already configure Firefox's file handling behavior for each type separately (I have, for example, previously set PDF files to prompt me to "Save as...", in order to prevent exploits from triggering when I click on a link, by preventing Firefox from opening them in-browser).
There is no reason I shouldn't be able to do what I have said I would like to. I should not HAVE to limit the number of max server connections in order to ACHIEVE filling a simple download queue with a 1-file-at-a-time limit.

In your Keychain under 'login' delete the VeriSign certificates and then quit and restart all browsers/itunes/app store.
http://apple.stackexchange.com/questions/180570/invalid-certificate-after-securi ty-update-2015-004-in-mavericks

Increasing max-streaming-connections-per-session has slow acknowledge response?

Our application is a Flex GUI with a WebLogic Server (BlaseDS) on a private network. We were originally using IE 6, but have upgraded to IE 8.
I am trying to use publish/subscribe messaging to monitor lengthy processes on the server and received incremental data. With 1 such process everything works fine. But we want to allow the user to subscribe to more than 1 message destination. So I increased the "max-streaming-connections-per-session" (default is 1) in the services-config.xml file
     <channel-definition id="process-notification-streaming-amf"
          class="mx.messaging.channels.StreamingAMFChannel">
          <endpoint url=https://{server.name}:{server.port}/{context.root}/messagebroker/streamingnotificationamf"
          class="flex.messaging.endpoints.StreamingAMFEndpoint"/>
          <properties>
               <user-agent-settings>
                    <user-agent match-on="MSIE" kickstart-bytes="2048"
                         max-streaming-connections-per-session="3" />
               </user-agent-settings>
          </properties>
     </channel-definition>
If we leave max-streaming-connections-per-session as the default value of 1 and try to subscribe to another message destination we get an error indicating limit has been reached:
     [BlaseDS]Endpoint with id 'process-notification-streaming-amf' cannot grant streaming connection to FlexClient with id '7FFC82DE-etc ' because max-streaming-connections-per-session limit of '1' has been reached.
     We upgraded to IE8 as documentation indicates IE8 allows for an increase of max-streaming-connections-per-session, where IE 6 is limited to 1. But increasing max-streaming-connections-per-session does not quite solve the problem. We have 3 consumers; consumer1, consumer2, consumer3. For each of these consumers, we add event listeners for MessageAckEvent.ACKNOWLEDGE and MessageEvent.MESSAGE.
     We call consumer1.subscribe(). When we receive the acknowledge message, we call consumer2.subscribe() (likewise with consumer3)
     The problem is it takes over 2 minutes to receive the acknowledge message from the call to consumer1.subscribe(). (With max-streaming-connections-per-session set to 1, the acknowledge message is received in a few seconds.)
     So, increasing max-streaming-connections-per-session removes the error about reaching a limit, but it appears to come with a cost of a big delay in a long delay on the call to subscribe? Or is there something we are missing?

I guess I will answer my own question. Hopefully this will be useful to someone else in the future...
The problem was coming from IE being limited to 1 connection by the registry. The solution can be found at:
http://support.microsoft.com/kb/282402
I manually performed the steps to update the registry, though microsoft provides a "Fix It"; MicrosoftFixit50098.msi
One other key element was to make sure to have kickstart-bytes="2048".

Max streaming connections per session error

I have a flex application that uses messaging with a streaming AMF connection, falling back to polling. When the max number of streaming connections on the server is reached, it does fall back to polling (at least it prints the max-streaming-clients error but the client connects, so I assume it is falling back - how can I tell?). However, occasionally the streaming connection will not initialize and it does not fall back - no messages are received on the client. The following error is logged on the server:
[EMST]09/25/2008 13:43:18.231 [ERROR] Endpoint with id 'my-streaming-amf' cannot grant streaming connection to FlexClient with id 'D5B8E3A1-1A1C-063E-84A6-6A743A1E4EE0' because max-streaming-connections-per-session limit of '1' has been reached.
This would make sense if the issue was caused by trying to initialize the streaming connection in two tabs of a browser, but I am only trying to initialize in one tab. Closing the browser (and thus destroying the session) does not fix it. The only solution I've found is to reboot the client machine. This has happened in both FireFox 3.0.2 and IE 7.
(1) What could cause the client to get in this state?
(2) When it happens, why doesn't it fall back to polling? Is the fallback only for when the server max connections is reached? When the streaming connection doesn't initialize, no messages are received.
(3) Is there a way to explicitly close the streaming connection on the client so we can fix this without rebooting?
Thanks!

Hi Mary. If you turn on Debug level logging on the client and the server you should be able to tell if you have fallen back to a polling channel after the attempt to connect over the streaming channel has been rejected. In the client log, you will see the flex application sending poll requests to the server at the polling interval configured in the channel and in the server log you should see that the server is receiving these requests.
The behaviour you are seeing seems very strange to me. The reason we have the max-streaming-connections-per-session limit on the server is because most browsers limit the number of active connections that can be made to a server from a single session. In IE for example, this is 2. What happens in most cases when the browser's connection limit is reached is that new connections are put on hold until one of the existing connections closes. This would cause your flex application to hang with no errors being reported on the client or the server. This is why we need the max-streaming-connections-per-session setting on the server. This prevents more than one persistent connection from being made from the same session, so the browser should never reach it's max connections per server limit and lock up.
It looks like you are somehow getting the browser to lock up even though the server is only limiting you to one streaming connection per session. It may be possible to do this if you reload the flex application in the browser (by doing a page refresh) in which case the browser could possibly briefly leave the streaming connection open in the background and when you tried to create a new streaming connection, the browser's connection limit to the server would have been reached and the application could hang. When the application hangs are you reloading the swf/page in the browser?
I really don't know why closing the browser wouldn't fix the problem. You're right that closing the browser should end the session. If you launch a new browser and load the swf do you get the same "cannot grant streaming connection" error on the server or is the browser just locked up, ie. no error is received on the client and the server?
You're not using a proxy server or anything like that are you that might be holding a connection open to the server?
-Alex

Max-streaming-connections-per-session limit

Hello,
i'm trying BlazeDS with Air app.
I set blazeDs on a Jboss Server with JMS adapter.
I configure it with a streamingAMF channel.
In user agent configuration i put msie, firefox value to 10 for the max-streaming-connections-per-session limit param.
In Air client configuration i instantiate a producer and a consumer on the same streaming AMF channel.
After the consumer.subscribe() i launch the producer.connect() on the server i get this error :
14:25:20,015 INFO [STDOUT] [BlazeDS] Endpoint with id 'my-streaming-amf' cannot grant streaming connection to FlexClient with id '497031A2-7B0D-019A-0E1D-7622A-A631D28' because max-streaming-connections-per-session limit of '1' has been reached.
Is it a limitation of use of blazeDs with Air app ?

First, if you change the limit to 10, then it should be 10 and not 1. If you still see 1, then please log a bug.
Second, you really want this limit to be 1 in IE and 4 in Firefox 2. But there's no reason to have more than 1 streaming connection from the server to the client unless you need to talk to two different endpoints.

Network error could not connect to server 127.0.0.1:5900

I have created a virtual machine in VM manager from a template and when I try to view the console I get the error "network error could not connect to server 127.0.0.1:5900"
The way that I have everything set up is this.... I am using Oracle VirtualBox installed on Windows 7 and I have created a virtual machine in VirtualBox with Oracle Virtual Server installed and I have Oracle VM Manager installed on that same virtual machine. After reading some forums I found out that I needed to install TightVNC to be able to use the console from Windows IE. So, I have installed TightVNC on my OVS but I can not even tell if it is running. I have also made sure that in iptables that port 5900 is open. I have confirmed this by using nmap to scan my ports and port 5900 is open for vnc when my virtual machine in vm manager is running. I just can not figure out why I am getting the network error unless TighVNC is not running on my OVS. Does anyone have any ideas for why this might be happeneing. Thanks

Hi,
it seemed when you registered the OVM server to the OVM manager, it did not deliver the correct hostname, but the local address instead.
See if your /etc/hosts contains the hostname in the 127.0.0.1 line.
Correct the /etc/hosts, remove the OVS server and reregister with OVM manager.
Regards
Sebastian

The ol' "mail.app holds 3 persistent connection per account" problem.

I've seen posts from a few years ago regarding this problem. My mail host only allows a certain number of persistent connection... ok, that's my bad for cheaping out on a host, I can dig that.
But there has to be an automated way to check, and disconnect for these types of situations. It doesn't seem right to have to change isp's when mail.app could have a little checkbox that says:
"don't persist IMAP connection because of ghetto ISP".
The fraction of a second it would take to reconnect, check and retrieve headers, then disconnect is something I can live with.
Does anyone have a workaround or know of a client that allows non-persistant imap connections?
Maybe an IMAP proxy daemon?
Any help will be rewarded by receiving the praise of all the other poor sods in the same boat. Your name will be rejoiced the world over, you will be a GOD!
Thanks

If Mail complains that it timed out on port 0, that means there is something wrong with your network and/or account settings.
Go to Apple Menu > System Preferences > Network, choose Network Port Configurations from the Show popup menu, and make sure that the configuration used to connect to Internet appears at the top of the list.
Also, try using a different method to connect to Internet, if possible, or connecting the computer to Internet as directly as possible, i.e. bypassing any routers that might be present, using an ethernet cable instead of wireless, etc., and see whether that makes a difference.
As to the logs, Mail doesn’t make it easy to get them. Open /Applications/Utilities/Console and do either File > Open Console Log, or File > Open System Log, or both, so that Console displays the contents of both system.log and console.log.
Apparently, you can tell Mail to provide detailed log information as described in the following article (I haven’t tried it myself):
Enable detailed Mail.app logging options
Also, Mail comes with a Turn On Logging AppleScript that I haven’t tried either, but that was discussed in the following thread:
MAIL LOG???

"Connect to Server" using a different user

Hi,
When I want to connect to a share on another Mac on my local network I use the Finder's "Go->Connect to Server" command.
I am always prompted for a username/password when I do this. Is there a way to cache this information somehow so it always knows which account info I want to use?
Thanks!
BT

You have the option to add the password to your keychain.
Just check the 'remember password in keychain' option.

Safari: network.http.max-connections-per-serve

Hello!
Does Safari have this kind of configuration parameter to set? (...as Firefox does)
We faced a problem on the 3.2 Windows version that we think might be connected to this parameter: in a web-page we have a lot of resources to include: CSS-es, JS-es and images, and, some of the CSS-es included in the <head> (with <link>s) were loaded later than in all other browsers (IE7, FF3, O9), and, in fact later than we would normally expect them.
We solved our problem by concatenating two of our CSS-es, but, anyway it would be useful to know Safari's limits and how to configure them.
Thanks!
P.S.: Yes, you are probably right thinking that this question should be posted on the "Safari for Windows" forum, but it seems to me that it is applicable to both versions and I would probably get a larger audience here.

See:
*http://hg.mozilla.org/mozilla-central/rev/d3ba16c3c8a0 (Bug 770331)

Exchange active sync cannot connect to server after migration mailbox user from exchange 2007 to exchange 2013 coexistence

Hello, everyone, my name is rafl
I have a problem with exchange 2013 active sync.
I have installed exchange 2013 coexistence with legacy exchange 2007, I have to migrate user mailboxes: [email protected] from exchange 2007 to exchange 2013.
but any problem with active sync connection on the mobile device after moving mailbox user. I reconfigure the exchange ActiveSync external connection domain (latest.domain.com) on mobile device replacing legacy exchange ActiveSync external connection domain
(legacy.domain.com)
the process of moving mailboxes successfully without error.
Access OWA for exchange 2007 and exchange 2013 is running normally
access mail from Outlook running normally
Certificate request has been installed and has no problem with it
The OWA virtual directory is configured for internal and external connections and different from the legacy exchange
The autodiscover virtual directory is configured for internal and external connections and different from the legacy exchange
ActiveSync virtual directory is configured for internal and external connections and different from the legacy exchange
user mailboxes are still on exchange 2007 is not problematic.
only problem with Exchange Active Sync on mobile devices, where I set up an email with android, iphone, windows phone. the error message: cannot connect to the server.
but, if I create a new user and create user mailboxes directly in exchange server 2013, ActiveSync can run without error on mobile device, access through OWA, MsOutlook, Outlook Anywhere also run normally.
only the results of user migration from exchange 2007 to exchange 2013 which is troubled with exchange ActiveSync connection.
any ideas for this problem, and what should I check on the exchange server ..?

i have run the activesync test connectivity and get some error :
Testing TCP port 443 on host domain.co.id to ensure it's listening and open.
The specified port is either blocked, not listening, or not producing the expected response.
A network error occurred while communicating with the remote host.
Elapsed Time: 3091 ms.
Testing TCP port 443 on host autodiscover.domain.co.id to ensure it's listening and open.
The specified port is either blocked, not listening, or not producing the expected response.
A network error occurred while communicating with the remote host.
Elapsed Time: 21072 ms.
Testing TCP port 80 on host autodiscover.domain.co.id to ensure it's listening and open.
The specified port is either blocked, not listening, or not producing the expected response.
A network error occurred while communicating with the remote host.
Elapsed Time: 21049 ms.
I have allowed access to port 443 (https) and 80 (http) on the firewall and re-run testconnectivity, but still with the same results. if I enable active sync for users who created directly in Exch 2013 there is no problem with the ActiveSync, just a problem
for users who moved from Exch 2007 to Exch 2013. @Android, iPhone, and Blackberry the error message "cannot connect to the server"

Software Update server keeps downloading updates when set not to

When I first setup our server (10.5.x) for software update, I looked over the option to to uncheck the box for automatically downloading updates from Apple. Since then, it has downloaded pretty much everything. I've unchecked that option and also selected it to delete old updates. It keeps downloading updates and won't delete older updates. We've since upgraded to 10.6.2. Same thing. I've even manually deleted ALL of the updates only to watch them all come back. Any clues?

Welcome to the wonderful world that is the Software Update Service
I have the same problem and have yet to find a sensible answer to this problem. I have three Macs at home and only need a small proportion of the updates that Apple supply yet no matter what I do, I can't seem to stop SUS downloading everything even though I don't want it and have told it not to download.
Best of luck getting any sense out of Apple on this. I'm going to try a home made solution to get round this problem.

Server keep running after user log-off?

Just need confirmation...
Does it keep running all services (mail, web, etc...) after logging off a user from macosx 10.5.4 server?

thanks. just need to create new user and log in because current user's finder is relaunching on almost every action with it... want to try if it's problem with user or all system.

Ipod Touch will not charge or connect to Itune, keeps reverting to apple logo?

My Ipod Touch will not charge or connect to Itunes/computer
Ipod switchs on charge logo appears then switchs to apple logo

Try:
- iOS: Not responding or does not turn on
- If not successful and you can't fully turn the iPod fully off, let the battery fully drain. After charging for an least an hour try the above again.
- If still not successful that indicates a hardware problem and an appointment at the Genius Bar of an Apple store is in order.

How can I use firefox between windows 7 & windows XP on a network, where we are using Samba server. If a user logs on different workstation(and OS) with same account.

we have a Linux fileserver with Samba share and have several users.
The users tend to be in different locations and need to login at different workstations, sometimes the workstation may be WindowsXP and othertimes it is windows7.
How can we use Firefox & Thunderbird in this situation.
I find that Windows7 uses compressed files to store the location information.
when a user logs on to win7 it creates a new account.

we have a Linux fileserver with Samba share and have several users.
The users tend to be in different locations and need to login at different workstations, sometimes the workstation may be WindowsXP and othertimes it is windows7.
How can we use Firefox & Thunderbird in this situation.
I find that Windows7 uses compressed files to store the location information.
when a user logs on to win7 it creates a new account.

There appears to be a persistent connection limit of 2 for cross domain requests

I am building a service that allows anybody to add real-time push updates to their website or web application. To do this I'm using the Access-Control-Allow-Origin header response from a cloud hosted server.
When establishing a connection we create one persistently connected HTTP connection and use a second connection to send requests (subscriptions/commands).
The problem I'm seeing looks very similar to the old 2 socket to a server connection limit (related to about:config "network.http.max-persistent-connections-per-server").
Is there a limitation on the number of persistent cross domain connections that can be established from a single Firefox process? If so, where is the configuration options to override this?
This problem can be easily duplicated by going to the following page in two tabs within the same Firefox process:
http://kwwika.com/Standalone/Demos/WorldCup2010/
You will notice that one page connects and starts to receive updates. The second will not connect until you close the first tab/page.
I'd like to be able to provide feedback to the users of Kwwika about this problem so any information you can provide me with would be very much appreciated.
== URL of affected sites ==
http://kwwika.com/Standalone/Demos/WorldCup2010/
== User Agent ==
Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/533.4 (KHTML, like Gecko) Chrome/5.0.375.70 Safari/533.4

[bump]

Network.http.max-persistent-connections-per-server keeps reverting to user set 4

Similar Messages

Maybe you are looking for