Bizarre RMI behavior

hi.
i have a very bizarre anomaly occurring and i am at a loss on where to even begin looking.
i have a client-server application that uses rmi for the network communication. the server is an app that listens for inbound tcp connections and then looks for a registered client that can service the request. clients obtain a remote reference to the server that they use to register a reference to themselves. they also use the remote server ref to put themselves in various availability states.
the server uses the remote client references to signal appropriate clients of some pending work.
the server uses a background 'ping' thread that periodically invokes a lightweight method on each client reference to ping each remote client. if this 'ping' doesn't return in a timely fashion (2sec) the remote reference is purged and the client must re-register with the server.
everything works perfectly for the most part except occasionally the 'ping' will timeout. some clients reside in the same bldg as the server yet others reside in a different state half way across the country. on average, the round-trip time for a 'ping' is around 120ms for remote clients and less than 50ms for clients local to the server.
what i usually find is that the timeout occurs in batches - usually all the remote clients will timeout but not the local suggesting a network issue. however, occasionally a group of remotes will timeout, all the locals will be ok, the next remote will be ok but the last remote will timeout... other times a local client will timeout...
i might add that this server is running on linux and the clients are all running on windows xp. i have another identical client-server setup that is identical except that all these clients also run on linux. at first, my app exhibited the exact same issues with timeouts on the linux-linux system. i searched the java bug database and discovered the issue with reverse dns doing a lookup when it creates an InetAddress. i then configured all of my linux clients to not use dns and the problem went away and i have yet to experience a single timeout since.
i'm blaming it on the network but my network engineer is blaming it on my app. he's run various diags on the network and everything checks out ok (ICMP ping times are fine during the timeout). maybe i should just blame it on windows!
i've tried turning on various rmi logging but there's so much junk in the log i don't know what to look for.
any help would be greatly appreciated!

thanks ejp for the reply.
it turns out that on the windows platform i need dns - i do not on the linux side. i have added the internal ips that the jvm system may be trying to reverse lookup to the the dns server - it made no difference. i then added all the ips to the etc/hosts file of the client - again no help. the server is configured to not use dns. i'm not entirely convinced it is dns - i was at first but if it were then these changes should have solved it - no?
i agree with you that 2 seconds may be a tad on the short side for a timeout across state boundaries - originally this was to all be in the same bldg so 2s was an eternity.
-ten6dsix

Similar Messages

Bizarre tabs behavior

After the 10.4.11 update Safari has had very bizarre tabs behavior.
http://www.99.se/attachments/safari/14197d1195126743-safari-3-bugg-bild-2.png
http://www.99.se/attachments/safari/14211d1195172108-safari-3-bugg-bild-1.png
http://www.99.se/attachments/safari/14212d1195172234-safari-3-bugg-bild-5.png
The active image isn't showing, or shows up completely screwed up. I've tried reverting to Safari 2, which fixed the problem, but when reapplying the update the problem comes back. I've tried every magic script I know to repair the system.
I have no plug-ins whatsoever.
Any ideas?

Sorry, no ideas, but I have an odd situation, too. When working with multiple tabs, I click a link that opens a new window, but it doesn't open a new window. It opens in oldest tab with the header (and URL) of the old web page, but the body being the pop-up.
Go figure that one out!

Can't add artwork - bizarre iTunes behavior

Okay - this is bizarre - I imported an album into iTunes 9 (downloaded from the Artist's website). I added the artwork and changed the genre and number of tracks, but these changes only "stuck" to about 4 of the 19 tracks. No matter what I did, I couldn't add the artwork to the 15 tracks that needed it. And - interestingly - if I next/previoused through the tracks on the album, all of the values that I changed would revert to the original values the tracks had when they imported. It was almost as if iTunes couldn't change the ID3 tags in the mp3 files. (Well, I guess it really couldn't when it came down to it.)
I solved the problem by right-clicking all the tracks and selecting "Create MP3 verssion" - essentially duplicating the MP3 files. I was then able to successfully get my changes to "stick" to the new files. (And then I deleted the originals I imported.)
What is this strange strange behavior all about?

There are three common causes for this:
your media are in a format that doesn't accommodate embedded artwork - typically WAV files. To embed artwork you'll need to convert to another format - Apple Lossless or AIFF if you want to preserve lossless quality, AAC or MP3 otherwise.
your media files are read only - to fix this, use Windows Explorer to find the folder that contains your files, right-click and select Properties. On the General tab there's a check box labeled "Read-only" - if this is checked, or is grey (sometimes blue), click the box so that the flag is unchecked and the box is white. Click OK, and then OK again when the "Apply changes to this folder, subfolders and files" option selected. Now try adding the artwork again.
Windows permissions issues are preventing iTunes from updating your media files (there's some anecdotal evidence of a change in this behavior in iTunes 12). See turingtest2's notes on Repair security permissions for iTunes for Windows for advice on fixing this, then try adding the artwork again.

Bizarre app behavior after waking from Sleep Mode

My iMac Core Duo runs like a top. It is fast and quiet. I have no problems with applications, both Universal and PowerPC. I maintain it with everything that I suggest others do in posts.
But, there is this one strange quirkiness that I have put off mentioning because I kept hoping that each new OS X update would address the issue. It has to do with waking my Mac from sleep mode with it logged into my user account. Not every time, but most of the time, if I put my Mac to sleep or it goes to sleep from inactivity while it is logged into my user account, I experience strange application behaviors upon waking.
For example, upon awaking Mail.app will contact my .Mac account and download any new mail that has arrived. But, if I try to open any of this mail, instead of opening, Mail.app will open a new outgoing email addressed to whomever sent me the mail I tried to open. Quitting and restarting does not solve this problem. Other apps demonstrate similar bizarre behavior. To only way to solve the issue is to log out and back into my user account.
However, if I log out of my user account and then put the Mac to sleep from the Login Manager I don't experience this issue. So at night I always log out before putting the Mac to sleep.
Anyone else experiencing similar issues? Any ideas on solving this? Do you think it is a bug in Universal Mac OS X?

Sorry this isn't a solution, rather is a commiseration. Often after my iMac wakes from sleep mode it refuses to see my internet connection. I have to unplug and reset my DSL modem to force the computer to see that it exists.
It sounds like the problem might be related, does anyone have any brilliant suggestions?
Thanks,
Lonnie

Bizarre email behavior

After having successful Mail use for years I now have a strange problem. Mail fails to send my emails using my Cox ISP. I receive emails without any problems. I can send my emails through my gmail account. Here is where it gets weird. I did all of the usual remedies, resetting the SMC, resetting PRAM, and performing verify disk and repairing disk permissions without correcting Mail's behavior. This behavior continued about one more week then corrected itself and I regained Mail send through my Cox account.
After being 'nice' for about 12 days, the problem has returned and I can send only by my gmail account. Any thoughts or suggestions?
Another point of information. My iPad (iOS 5.1.1) Mail sends and receives email OK through my Cox ISP but I always get the message "The mail server 'pop.west.cox.net' is not responding. Verify that you have entered the correct account info in Mail settings." Since I get and send all email without any problems, I just sigh and ignore it.
All of my device accounts have identical account info. I now have iFatigue.
On a better note, my iPhone (iOS 6.0.1) Mail performs properly as expected.

I wish I could help but I still live with the problem on my iMac. My iPhone (4S) has no problem but my iPad would not recognize the Cox accounts. I finally deleted the Cox email accounts and re-entered them. That worked for the iPad. Bizarre!
On the iMac, I apparently at random, I successful send email using either of my Cox email addresses, but usually Mail forces me choose my Gmail account to send. Again, bizarre!

Why is 2.1.3 causing bizarre adjustment behavior?

I just downloaded and installed 2.1.3. Soon after I tried to adjust an image; but as soon as I moved the shadow slider the image went all wild — white with green lines. I tried another image, and when I moved the highlight slider the same thing happened, and same with contrast. (No problems on PS.) These images are on a Compact Flash card previously downloaded where other images behaved normally during adjustment, so it is not card corruption. (Camera is Nikon D300)
Is this a known issue on v 2.1.3? Or is there something idiosyncratic going on here? I would be grateful for any advice. I do minimal adjustments and prefer to stay in Aperture.
Thanks, Susan

This is my first use of this thing and I can't figure out how to ask a question, so I'm not replying I'm
asking. I am experiencing bizarre responses to various sliders in the adjustment inspector, shadows,
contrast and others. I have had no problems like this since I got the system last June, importing only
raw digital images. However for the last two weeks I have been scanning b&w negatives at 6400 dpi
as tiff files, which it turns out are at least 60MB and often 120MB in size. All of the Aperture problems started since this. I do not understand this but I'm wondering if the size of the files is something
Aperture is not comfortable with? I'm getting half black or sudden all black screens, occasionally
colored constructions, and complete shutdowns requiring relaunching three or four times every couple of hours. Any response would be helpful. I really like Aperture and hope to solve this. Thank you.greg

Bizarre mouse behavior on rollOut - bug??

I built a custom cursor out of a swf file and I am applying
it on rollover to a specific item. That part works fine. The
problem is the on rollOut event (or mouseOut I've tried both same
exact behavior) fires CONTINUOUSLY when I move the mouse to the
RIGHT. It only fires when I move the mouse to the right. Let me
give an example:
app starts, normal mouse cursor
i roll over an item, swf cursor applied
normal behavior is for swf cursor to be applied over the
item, and off when rolled out of it.
if i move my mouse to the right while still OVER the item,
the event fires for every pixel i move even though i haven't rolled
out.
if i leave the cursor on and move off the item and move
around in the application, the event still fires every single pixel
i move to the right. not only does it fire - it fires a rollout
event up the entire chain of my application. it fires an on rollout
event for every single component in the application including the
APPLICATION itself!!
i'm going nuts on this one, anyone have some insight? the swf
cursor is nothing fancy, just a bitmap, text, and one small little
animation with no scripts.

The issue seems to be resolved (after a longgggg time of
debugging) to be related the the X and Y offset parameters of the
setCursor method. My cursor was designed to have no offset so i
left it blank. I tried 0,0 but that didn't fix the problem.
Putting in any number (1,1) seems to fix the problem. this
must be a bug of some sort in the constructor for this class. Can
any of the Mods/Devs mark this as a bug to investigate?

Bizarre flash behavior browser vs projector in Firefox

I have a rather strange problem and it looks to be browser
dependant. The project was made for player 8 with action script 1.0
and is linked below. When viewing this project as an exe file, the
behavior is exactly how it should be, however when viewing it
through a browser via firefox, the behavior is completely
different. Flash will only pick up one click at a timem meaning for
every button or link I click on, on the swf, I have to first click
outside the flash document to change the focus back and forth to
the swf otherwise the input is ignored. This behavior is not
present in IE. The first click is ignored in IE but once flash has
the focus, it works as it should from then on. IE was tested with 6
and 7. Firefox version 1.5
The zipped exe file
http://www.hesaa.org/TestPages/tcnj
- 2007 v3f.zip
The embed swf file
http://www.hesaa.org/TestPages/TCNJ
- 2007.html
Update: I've updated to flash plugin 9, no change. I've also
tried this same file on Firefox 2.0 and it works without any
problems, so the problem must be related to firefox 1.5

Hi James,
Thank you for sharing your solution with us. It will help other users with a similar issue.
Thanks,
Preran

Strange RMI behavior and data

Hello:
I have a very difficult and strange problem. I am hoping somebody can spot what is going on or maybe tell me I am way off.
I have a client-server application that uses RMI for network communication. The server is a database and the clients insert, update and delete data from it. When a client starts up, it does Naming.lookup, and upon a success, will register some callback methods so that it gets notified of any changes to the database.
Everything works, but the performance is somewhat sketchy. For example, I have one test where one client sends over 1,000 new items to the database. After about 10 seconds of great performance it suddenly stops, almost like it lost its connection. Then, after about 10-30 seconds later, it starts right back up and runs for another 10 seconds. I am not getting any exceptions thrown.
At first I thought this was a network configuration issue, but I am able to resolve names and IP address backwards and forwards across all the clients and the server.
I installed ethereal, a network sniffer, to try to find out what was going wrong. I noticed a strange pattern, between my clients and server. Every 10 seconds there is strange exchange between each client and the server that last about .1 of a second.
The client always sends the same 41 bytes of data and the server sends the same 23 bytes back, except for the second-to-last byte. That one is always incremented.
The exchange is continuously happening even when none of the clients are doing anything, which leads me to believe that it is just a normal part of RMI and isn�t doing anything. However, I have also noticed that my slowdowns ALWAYS occur when this happens. Below is a snippet of the data exchange between one of my clients and the server.
Has anybody ever seen this, or does anybody here have an idea of what is going wrong?
Client:
00000000 52
Server:
00000000 53
Client:
00000001 50 ac ed 00 05 77 22 00 00 00 00 00 00 00 02 01
00000011 82 0d da 00 00 01 0c ea 55 13 9f 80 00 ff ff ff
00000021 ff c6 90 44 fb 34 76 b1 8d
Server:
00000001 51 ac ed 00 05 77 10 01 01 82 0d da 00 00 01 0c
00000011 ea 55 13 9f 85 17 01
Client:
0000002A 52
Server:
00000018 53
Client:
0000002B 50 ac ed 00 05 77 22 00 00 00 00 00 00 00 03 01
0000003B 82 0d da 00 00 01 0c ea 55 13 9f 80 00 ff ff ff
0000004B ff c6 90 44 fb 34 76 b1 8d
Server:
00000019 51 ac ed 00 05 77 10 01 01 82 0d da 00 00 01 0c
00000029 ea 55 13 9f 85 18 01

Thanks for everybody�s help so far.
I have done some more testing and I found some usefull information in the transport.tcp log on the server side.
First, some background: I am now running with a single client attached to the database. My client is at 192.168.0.63 and my server at 192.168.0.21.
When my program starts up, my client opens up an RMI connection with the database to get the current data and to send database changes one. The client port is 3110 and the server port is 58252.
After I make the first change, the server makes a RMI connection back to my client to sends it datbase update. The server port is 58547 and the client is 3087.
I am sending the same two updates to the database every 5 seconds or so. After about 30 - 60, the database takes 12 seconds to contact the client where it normally take 0.005 seconds.
Just after the previous successfull update I saw this message
06:33 sun.rmi.transport.TCPChannel$Reaper run
FINER: RMI ConnectionExpiration-[192.168.0.63:3087]: wake up
Then at 6:36 I attemped to re-save another database object, I got my message that the server was about to attempt to fire off another update to the clients. Immediatly after that I got this log message
06:36 run.rmi.transport.tcp.TCPChannel new connection
FINE: Thread 3 - reuse connection
6 seconds afterwards I get this
6:44 sun.rmi.transport..tcp.TCPConnectionHandler run
FINE: RMI TCP Connection(3)-192.168.0.63 accepted from 192.168.0.63:3201
6: 44 sun.rmi.transport..tcp.TCPConnectionHandler run
FINE: RMI TCP Connection(3)-192.168.0.63(58252) suggesting 192.168.0.63:3201
6: 44 sun.rmi.transport..tcp.TCPConnectionHandler run
FINE: RMI TCP Connection(3)-192.168.0.63(58252) using 192.168.0.63:3087
The update to the client finishes at 6:48
In the packet sniffer I saw the same 12 second delay between the ACK and the PSH,ACK packets. In that delay, I saw that an RMI connection was initiated between the client on port 3201 and the server on port 585252. I am assuming that that connection is what was showing up at 6:44.
I guess that the connection between the database and the listener on the client is expiring for some reason. What I don't understand is why and who terminated the connection, the client or the server?
I was able to remove the program that was hiding my transport.tcp log messages, but I didn't see any of the connection re-establishment on the client side. Maybe I just didn't look hard enough

Bizarre CS6 behavior

Recently my CS6 has exhibited a weird behavior--when I open a jpeg, for example, only half of the file is displayed--the other half appears as a blank area in the file display area. When I click on the blank space, the rest of the file will appear. Windows 7 up to date, 9 GB Ram, i7 processor. This just started a couple of weeks ago and I have no idea what's up. I've been considering uninstalling and reinstalling but fear for my plugins and understand that reinstalls are not straightforward. Any ideas???
Below is a screenshot cexample

Mylenium wrote:
or turn off OpenGL in the prefs.
Just for reference, it's not called "OpenGL" in the prefs with Photoshop CS6. You might want to change the advice to the following:
... or turn off "Use Graphics Processor" in the prefs.
-Noel

Bizarre ftp behavior in DW8-- check this out

I'll preface this by specifying that while using MX2004, I
had no problems with ftp whatsoever; I maintain a massive,
constantly-changing, image-intensive site and I simply do not have
time to select updated/added files one-by-one to upload; with
MX2004, I could simply select the main "web" folder, hit "put", and
all modified/new files would be uploaded automatically..........not
so in DW8:
when I try to use "select newer local files," ususally a box
pops up referencing a temp file ("putting temp2342.htm"), and
either nothing happens, or my ENTIRE site, with all subdirectories
is selected for upload; now, I've tried as many different things as
I can think of in terms of modifying settings in the site
definition, and I've noticed the following: if I modify the remote
root directory in the site definition settings, and then attempt a
"select newer local files," it will actually work; i.e. it will
actually scan for newer files, and select new/updated files
correctly; however, if I then try to do this a second time, nothing
will happen, like before;
each time I change the remote directory in the settings,
however (say, to "/web" or even just "/"), it seems to work once,
and then stops; I can restart the program, etc. and the
functionality doesn't seem to return unless I change the setting
again and again........
if I can't get some sort of solution to this issue, I will
likely go back to MX2004, which I REALLY do not want to do;
***DW developers, please take note: from searching on the
web, I see that LOTS of people seem to have similar problems with
the revamped FTP functionality in DW8; for God's sakes, just
re-implement the old MX2004 ftp behavior...or something that WORKS
like before.

quote:
Originally posted by:
Newsgroup User
I'm trying to reproduce your behavior in DW8 on both the Mac
and the PC. I
cannot.
Can you give me an explicit set of steps that cause the
symptoms to occur
every time, please?
Does this happen on every site or just one?
Murray --- ICQ 71997575
Adobe Community Expert
(If you *MUST* email me, don't LAUGH when you do so!)
==================
ok, sorry for the much delayed response; I've been
experimenting with different settings, trying to get to the root--
no pun intended-- of the problem. It seems to me that this problem
with the "scan for newer files" function relates to my particular
ftp server config, and I can sometimes get it to work by modifying
the remote root directory.
Without specifying a remote root dir-- which I did not have
to do in MX2004--the ftp server displays four main folders: "cgi",
"files", "logs" and "web", with the "web" folder containing the
actual site files. On the local end, I usually specify a main
folder, eg. "main web site" that contains my templates directory,
and a corresponding "web" directory (with the local site files).
When a remote root directory is not specified, the "scan for newer"
function usually does not work. HOWEVER, if I specify "/web/" as
the remote root directory, and change my local root folder to
"/web" as well, the scan function DOES work.
I don't like this config, however, because my local site is
defined as being rooted in the local "web" folder, and thus, I
don't see my template directory, etc.
Here is the strange part, though: if I then go into the site
manager and change the remote root to "/" and change my local root
back to "main web site," the cache will be recreated, and the "scan
for newer" function with actually WORK. This is where it's finicky
though, because if I shut down the program and restart, the
scanning behavior will then revert to the previous behavior (cf.
initial post).
Something I've learned is that DW is unable to write to the
"base root" directory-- ie. "/"--but, of course, can write to the
"/web/" directory; so I'm thinking it having an issue creating a
timestamp, or something along those lines; however, this still
doesn't explain why it would intermittently work as specified in
the paragraph above.
Here's a question: is there a way for me to specify "/web/ as
the remote root directory, and maintain the "main web site" dir as
the local root folder (so I can see my templates folder, and the
underlying local "web" directory)? This seems to be problematic,
b/c if I try to upload with this config, DW wants to create another
"web" directory within the remote "web" as the local and remote
file listings are not on the same level, so to speak.
Sorry for the length, but I'm trying to supply as many
details as possible. Finally, I have not yet tried DW8 with my
other sites, so I'm not sure if it's unique to this site (I have a
feeling that it is, though); my hosting provider is XO
Communications, if that offers any insight into this
situation.......

Bizarre Speaker Behavior

My speaker sound has disappeared, unless I install a set of headphones into the microphone jack on the front of the computer, then it's all happy. The headphone jacks seem to work fine. I restored the computer to the factory-loaded condition, reloading Windows7 last night, but the problem is still there. There doesn't seem to be anything in the BIOS that can result in this behavior, and I'm stumped.
I have a dv7-3060US model and it's only 6 months old. Any help would be greatly appreciated.

I have the same issue with my Pavilion dv7 2022tx!
It really sucks and I am not happy with this as I have has this ever since I got the laptop....
Please if you can help me out, I would appreciate it.

Bizarre Sticky Behavior

Hi...
I was trying to save a sticky to my desktop and for some bizarre reason it was created under my Final Cut Express program...
I would open up the “open using” option and all were geared toward a video file... So that means the Sticky was “created” under some pseudo Final Cut Express “form” or another...
I even selected to have the Sticky “open under” it’s own Stick program and it was not recognized...
any suggestions, comments or opinions are appreciated
Thank you in advance...

Try this... another pref file com.apple.finder.plist
/Users/YourName/Library/Preferences. Same drill, drag that file from the Preferences folder to the Trash, empty the Trash and reboot.
If these issues continue, try downloading and installing the 10.5.6 combo update available here.
http://support.apple.com/downloads/MacOS_X_10_5_6_ComboUpdate
If you do this, run Disk Utility and Repair Disk Permissions before and after you install the update.
Launch Disk Utility. (Applications/Utilities) Select MacintoshHD in the panel on the left, select FirstAid, then click: Repair Disk Permissions. Quit DU when it's finished and reboot.
*"(I thought it would go away maybe from not using my computer, ha, ha)."* Stranger things have happened just leaving a computer turned off over night.
Carolyn

Bizarre function behavior

When I enter this formula "=IF(WEEKDAY(A3)=6,A33,A31)" in cell A4 (A3 contains a date and is a date format), A4 is blank, and every cell to it's right contains the expected result! When I click back into A4, there's no hint of a formula, and the cell is now formatted as text.
If I enter the exact same formula in cell D4, it works as expected. Very weird.
I've tried this multiple times, using the formula editor or simply using the entry bar. I've done it with existing files and new files.

Very surprising behavior which I can't reproduce.
Here I get:
In A4, the formula is:
=IF(WEEKDAY(A3)=6,A33,A31)
and I used filltoright.
I have an idea but it seems really silly.
Are you sure that there isn't a conditional formatting applied to the cell A4 ?
Yvan KOENIG (VALLAURIS, France) mardi 8 septembre 2009 22:46:51

RMI behavior in concurrent connections

hello guys!
I would like to know if a rmi application open more then one port if necessary when it need to handle many concurrent connection. Cause i've seen in JVM logs it handle a lot of threads but i don't know if in every thread it's the same port or it's create a new one.
I mean there's a entry point port for example 1099 but at low level JVM opens any other port. Is that thuth or i am completely lost
thanks for any comments
endyamon

RMI accepts multiple concurrent connections and despatches them all in separate threads.
TCP/IP gives all accepted sockets the same port number as the listening socket.
So the answer is yes and yes.

Bizarre RMI behavior

Similar Messages

Maybe you are looking for