As the title say i can't connect to the client. The OS is Lubuntu 12.04 64bit and the software is installed directly from the Synaptic package manager (It's x86). The software runs fine after the installation till today, when i could not connect to the client anymore. First thing i check is firewall:

$ sudo iptables -L

Chain INPUT (policy ACCEPT)
target     prot opt source               destination         

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination         

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

Checking the state of port

netstat -a | grep 31416

tcp        0      0 *:31416                 *:*                     LISTEN     
tcp        0      0 localhost:55667         localhost:31416         CLOSE_WAIT

Status of the client

$ sudo ./boinc-client status

 * Status of BOINC core client: running
 * Scheduling of BOINC core client: 4070
pid 4070's current scheduling policy: SCHED_OTHER
pid 4070's current scheduling priority: 0
 * OOM killer status for BOINC core client:
PID 4070: adj 0
, score 
5

It seems everything is fine with the client.Reinstalling the software with purge(deleting config files) does not help. Then i'm trying to find some clue in the log files. In the file:
time_stats_log i saw rows with

net_unknown

In the file:
client_state.xml

<ip_addr>127.0.1.1</ip_addr>

What should i do?

In the file:
client_state.xml

<ip_addr>127.0.1.1</ip_addr>

Why do you do this? Doesn't your computer have an IP address on the local network? As that's the address that should be in there, normally in the form of e.g. 192.168.1.1
And even then, you don't have to add that manually, as it'll be added by the BOINC client.

As for why the BOINC Manager cannot connect to the client, have you checked the most simple thing, that the client is actually running?

---

Why do you do this? Doesn't your computer have an IP address on the local network?

My computer have ip 192.168.1.104


I don't add anything to this file, i think the IP address needs to be 127.0.0.1. That's why i mentioned it. I can use the boinccmd, and the IP address is the same.

$ boinccmd --get_host_info

timezone: 10800
  domain name: nemirco
  IP addr: 127.0.1.1
  #CPUS: 2
  CPU vendor: GenuineIntel
  CPU model: Intel(R) Core(TM)2 Duo CPU     E8500  @ 3.16GHz [Family 6 Model 23 Stepping 6]
  CPU FP OPS: 1000000000.000000
  CPU int OPS: 1000000000.000000
  CPU mem BW: 1000000000.000000
  OS name: Linux
  OS version: 3.2.0-31-generic
  mem size: 4146380800.000000
  cache size: 6291456.000000
  swap size: 0.000000
  disk size: 105687318528.000000
  disk free: 20163424256.000000

As for why the BOINC Manager cannot connect to the client, have you checked the most simple thing, that the client is actually running?

I'm sure the client is up and running: This is the initialization messages:

$ boinccmd --get_messages

1: 24-Sep-2012 19:37:53 (low) [] Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu
2: 24-Sep-2012 19:37:53 (low) [] log flags: file_xfer, sched_ops, task
3: 24-Sep-2012 19:37:53 (low) [] Libraries: libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
4: 24-Sep-2012 19:37:53 (low) [] Data directory: /var/lib/boinc-client
5: 24-Sep-2012 19:37:53 (low) [] Processor: 2 GenuineIntel Intel(R) Core(TM)2 Duo CPU     E8500  @ 3.16GHz [Family 6 Model 23 Stepping 6]
6: 24-Sep-2012 19:37:53 (low) [] Processor: 6.00 MB cache
7: 24-Sep-2012 19:37:53 (low) [] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm sse4_1 lahf_lm dtherm tpr_shadow vnmi flexpriority
8: 24-Sep-2012 19:37:53 (low) [] OS: Linux: 3.2.0-31-generic
9: 24-Sep-2012 19:37:53 (low) [] Memory: 3.86 GB physical, 0 bytes virtual
10: 24-Sep-2012 19:37:53 (low) [] Disk: 98.43 GB total, 18.78 GB free
11: 24-Sep-2012 19:37:53 (low) [] Local time is UTC +3 hours
12: 24-Sep-2012 19:37:53 (low) [] VirtualBox version: 4.1.12_Ubuntur77245
13: 24-Sep-2012 19:37:53 (low) [] No usable GPUs found
14: 24-Sep-2012 19:37:53 (low) [] Config: GUI RPC allowed from:
15: 24-Sep-2012 19:37:53 (low) [] No general preferences found - using defaults
16: 24-Sep-2012 19:37:53 (low) [] Reading preferences override file
17: 24-Sep-2012 19:37:53 (low) [] Preferences:
18: 24-Sep-2012 19:37:53 (low) [] max memory usage when active: 1977.15MB
19: 24-Sep-2012 19:37:53 (low) [] max memory usage when idle: 3558.87MB
20: 24-Sep-2012 19:37:53 (low) [] max disk usage: 10.00GB
21: 24-Sep-2012 19:37:53 (low) [] don't use GPU while active
22: 24-Sep-2012 19:37:53 (low) [] suspend work if non-BOINC CPU load exceeds 25 %
23: 24-Sep-2012 19:37:53 (low) [] (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
24: 24-Sep-2012 19:37:53 (low) [] Not using a proxy
25: 24-Sep-2012 19:37:53 (low) [] This computer is not attached to any projects
26: 24-Sep-2012 19:37:53 (low) [] Visit http://boinc.berkeley.edu for instructions

No, the IP address in client_state.xml is the IP address of your computer, not the localhost default address 127.0.0.1

The BOINC binary and the BOINC Manager binary do talk to each other on TCP port 31416 on the localhost loopback address 127.0.0.1, but that doesn't mean that the <ip_address/> entry must be that value.

By the way, you run a 64bit version of BOINC, as shown by Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu.

Now, I'm sure it's none of these, but do check the options that normally generate the "Manager cannot connect to a client" message at this BOINC FAQ.

---

Now, I'm sure it's none of these, but do check the options that normally generate the "Manager cannot connect to a client" message at this BOINC FAQ.

Yes, I check them all including reinstall of fglrx video driver

At the moment i got a workaround:
Start manager > Advanced > Select computer... > enter in the host field 127.0.0.1
After that the manager get connected with client. But then the another problem has come - the command menus disappear, and i can't set the CPU loading percentage. I decide to post the image of my program window and starting to decreasing the size of it.And then menus appear again.It seems my system is buggy.

Thank you for giving me of your time Ageless!

At the moment i got a workaround:
Start manager > Advanced > Select computer... > enter in the host field 127.0.0.1

Can you check if there's anything in the gui_rpc_auth.cfg file?
There should be a randomly generated 32 character hexadecimal pass code in it, although you can change it for one of your own passwords. If you do, make sure to only add the word to the same line the 32 character pass code is on, don't add a carriage return or anything.

Or you can delete the gui_rpc_auth.cfg file and then restart BOINC. That may jar loose any stuck things as well.

Finally, you're not starting your boinc binary or boincmgr binary with any added attributes, are you? Or in case you are, what are they?

---

Can you check if there's anything in the gui_rpc_auth.cfg file?

The file was empty (got 1 byte length but there is no information in it), after deleting it, i can't connect to client , even with workaround.

Finally, you're not starting your boinc binary or boincmgr binary with any added attributes, are you? Or in case you are, what are they?

Everything is started without any attributes.

The file was empty (got 1 byte length but there is no information in it)

That ain't normal. I'm starting to wonder if it's a bug in the BOINC version, but since you got it from the Synaptic package manager, it's built by the software maintainer for that version of Lubuntu. It's not a Berkeley built version, and besides, if it were it would be 7.0.28 as that's the recommended version at this time.

Would you mind trying the Berkeley build version?
It does install to a different directory than the package maintained ones, though, so do take care of that.

---

Would you mind trying the Berkeley build version?

The situation is the same, I can connect only with my workaround.I just copy new binaries (boinc, boincmgr and boingccmd) to the place of the old ones.File (gui_rpc_auth.cfg) is there and it's empty.

By the way i tried something else:
Using my "Virtual Box" (virtualization software), I made a new installation of my OS (Lubuntu 12.04 64bit), update the system and install BOINC 7.0.27. Everything works fine! The file gui_rpc_auth.cfg is empty.

Another thing that I saw is in the down left corner of the
application windows. It says "Connected to 127.0.0.1 (7.0.28)"
on the virtual machine it says "Connected to localhost (7.0.27)"
I try to replace 127.0.0.1 with localhost in my workaround and when I press last letter (t) in the password area appear 32 dots like I'm entering some password. Anyway i can't connect to the client with this strange appearing password and without it.

Is it possible another gui_rpc_auth.cfg file is on your system, within a directory that BOINC can read?
Or is it possible that there's another BOINC installed on the system that's playing interference?
Sorry, I'm a Windows guy, I had something like that once when I tried to run a BOINC Manager not from the BOINC programs directory.

---

Is it possible another gui_rpc_auth.cfg file is on your system, within a directory that BOINC can read?

Yes i found 5 of them some empty or with some string in it. I just delete some of them and everything is fine now. Thanks for the idea.

Great! :-)

---

Same problem here. Ubuntu 12.04, repository 7.0.27, starting Boinc manager from a gui startup folder. Always, the manager cannot connect to core client. So I stop the manager and restart. Then it works. This tells me all my tcp/ip settings are OK and that the core client starts running.

I got tired of this routine. (I had left boinc manager in the startup folder as a reminder to get it running.) So I followed this advice:
"Is it possible another gui_rpc_auth.cfg file is on your system, within a directory that BOINC can read?"
"Yes i found 5 of them some empty or with some string in it. I just delete some of them and everything is fine now."

I found several gui_rpc_auth.cfg's and deleted them. Now the manager won't connect with the core client no matter what I try. Reinstalling didn't work, either.

Here's a screen print. My comments are in ALL CAPS:

craig@craig:~$ boinc & boincmgr &
[1] 4245
[2] 4246
craig@craig:~$ 16-Oct-2012 15:02:16 [---] No config file found - using defaults
BOINC MANAGER HAS NO WAY I CAN FIND TO SPECIFY WHERE TO LOOK FOR THE CONFIG FILE.

16-Oct-2012 15:02:16 [---] Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu
16-Oct-2012 15:02:16 [---] log flags: file_xfer, sched_ops, task
16-Oct-2012 15:02:16 [---] Libraries: libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
16-Oct-2012 15:02:16 [---] Data directory: /home/craig
BOINC MANAGER HAS NO WAY I CAN FIND TO SPECIFY THE USUAL DATA DIRECTORY IN /VAR

16-Oct-2012 15:02:16 [---] Processor: 2 AuthenticAMD AMD Athlon(tm) II X2 250 Processor [Family 16 Model 6 Stepping 3]
16-Oct-2012 15:02:16 [---] Processor: 1.00 MB cache
16-Oct-2012 15:02:16 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt npt lbrv svm_lock nrip_save
16-Oct-2012 15:02:16 [---] OS: Linux: 3.2.0-32-generic
16-Oct-2012 15:02:16 [---] Memory: 7.80 GB physical, 37.25 GB virtual
16-Oct-2012 15:02:16 [---] Disk: 385.02 GB total, 340.58 GB free
16-Oct-2012 15:02:16 [---] Local time is UTC -7 hours
16-Oct-2012 15:02:16 [---] VirtualBox version: 4.1.12_Ubuntur77245
16-Oct-2012 15:02:16 [---] ATI GPU 0: Cedar (CAL version 1.4.1720, 512MB, 432MB available, 208 GFLOPS peak)
16-Oct-2012 15:02:16 [---] OpenCL: ATI GPU 0: Cedar (driver version CAL 1.4.1720, device version OpenCL 1.2 AMD-APP (923.1), 512MB, 432MB available)
16-Oct-2012 15:02:16 [---] No general preferences found - using defaults
I HAVE SET PREFERENCES. WHERE IS BOINC MANAGER LOOKING?

16-Oct-2012 15:02:16 [---] Preferences:
16-Oct-2012 15:02:16 [---] max memory usage when active: 3993.28MB
16-Oct-2012 15:02:16 [---] max memory usage when idle: 7187.90MB
16-Oct-2012 15:02:16 [---] max disk usage: 10.00GB
16-Oct-2012 15:02:16 [---] don't use GPU while active
16-Oct-2012 15:02:16 [---] suspend work if non-BOINC CPU load exceeds 25 %
16-Oct-2012 15:02:16 [---] (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
dir_open: Could not open directory 'slots'.
16-Oct-2012 15:02:16 [---] Not using a proxy
16-Oct-2012 15:02:16 [---] This computer is not attached to any projects
THIS COMPUTER IS OF COURSE ATTACHED TO AT LEAST ONE PROJECT.

16-Oct-2012 15:02:16 [---] Visit http://boinc.berkeley.edu for instructions
Initialization completed

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.

(boincmgr:4246): LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.
boinc & boincmgr &16-Oct-2012 15:04:57 [---] Suspending computation - CPU is busy
16-Oct-2012 15:05:37 [---] Resuming computation

From the suspension and resumption of computation I assume the boinc client is working fine. Two questions:
1. How can I abandon the repository and switch to the Berkeley _without_losing_my_work?
2. Failing that, how can I help the manager locate the core client?

Same problem here. Ubuntu 12.04, repository 7.0.27, starting Boinc manager from a gui startup folder. Always, the manager cannot connect to core client. So I stop the manager and restart. Then it works. This tells me all my tcp/ip settings are OK and that the core client starts running.

For BOINC Manager to connect to a client, either (a) the client must be running already, or (b) the operating system must be in a suitable state for the client to start up when the Manager instructs it to.

If your manager fails to start the client when run automatically from a startup folder, but works after a stop/restart (presumably, after some little time has elapsed), that suggests to me that everything is in the right place and findable.

Some years ago, there were regular discussions here about inserting delays into Linux startup scripts: the idea being that BOINC had dependencies on other system components (network drivers? graphics X-drivers?), and couldn't start until the OS had fully initialised those components.

I'm not a Linux person, but Jord may be able to find and point you to those earlier discussions.

Inserting a time-pause at BOINC start so other stuff can load before BOINC starts doing things. Make a cc_config.xml file in your /etc/boinc-client/ directory, with in it:

<cc_config>
<options>
<start_delay>N</start_delay> 
</options>
</cc_config>

Where N is the amount of seconds you want BOINC to wait.
E.g. <start_delay>300</start_delay> will make BOINC wait 5 minutes, or 300 seconds, before it starts running applications.

The config file that BOINC says it didn't find, is the cc_config.xml file as above. By default this file isn't made until either you make it, or you add exceptions through BOINC Manager's advanced preferences menu. So it's normal for BOINC to report that it didn't find any such file, when you never made it.

BOINC MANAGER HAS NO WAY I CAN FIND TO SPECIFY THE USUAL DATA DIRECTORY IN /VAR

BOINC Manager is just a GUI, it doesn't do anything special.

The BOINC client will only put its data directory in /var/lib/boinc/ when you use the BOINC from your Linux repositories. The Berkeley installer will put everything in your home directory.

For more info see http://boinc.berkeley.edu/wiki/Installing_BOINC#The_Berkeley_Installer and http://boinc.berkeley.edu/wiki/Installing_BOINC_on_Ubuntu.

I HAVE SET PREFERENCES. WHERE IS BOINC MANAGER LOOKING?

Again, this is BOINC checking, not BOINC Manager.
It's checking in the directory where cc_config.xml, global_prefs.xml and gui_rpc_auth.cfg live in. I don't know what that is under your home directory, but in repositories it's in /etc/boinc-client/

THIS COMPUTER IS OF COURSE ATTACHED TO AT LEAST ONE PROJECT.

Perhaps under the other BOINC on your system it is, but not under this one. Look, in Windows and Mac OS X there's just one installer, which will always install to the same directories, as depicted by the Windows registry, or the configuration file in OS X.

With the Linux installers, you'll have to watch out which ones you used before, the Berkeley installer, or the one from repositories, as these aren't built by the same people and aren't installing to the same places.

As explained earlier, the Berkeley installer installs to your home directory.
The repositories installer:

1. Puts the BOINC binaries (boinc, boinccmd and boincmgr) in /usr/bin .
2. Creates a configuration directory /etc/boinc-client containing files cc_config.xml, global_prefs_override.xml, gui_rpc_auth.cfg, and remote_hosts.cfg (It is a Debian convention that such files live under /etc)
3. Creates the working directory /var/lib/boinc-client/ for BOINC data files and the slots and projects directories. Also creates links from this directory to the files in /etc/boinc-client
4. Creates a user named boinc. For improved security, this boinc user owns the BOINC working directory and all the data files and sub-directories created under the working directory.
5. Creates the daemon "init" script, /etc/init.d/boinc-client, and a set of configuration options for the daemon at /etc/default/boinc-client
6. Sets up the daemon to autostart at boot time.

1. How can I abandon the repository and switch to the Berkeley _without_losing_my_work?

Since they install to different locations, first go check in your home/ directory if you have anything BOINC related there. Then start the client from there. If you want to use the data from the repositories, you'll have to do some moving over of directory contents and at least the client_state.xml and client_state_prev.xml files. But that's too messy to go explain in this post. And I am still not sure which version you now run and if you have another on your system.

2. Failing that, how can I help the manager locate the core client?

Make sure that the client is actually running. ;-)

---

E.g. <start_delay>300</start_delay> will make BOINC wait 5 minutes, or 300 seconds, before it starts running applications.

I don't think that's what Richard meant. As I had to be told numerous times by JM7 ;-) the delay has to occur before the BOINC process itself starts running, not the project applications.

Gruß,
Gundolf

Well, in that case do a check for sleep: http://linux.about.com/library/cmd/blcmdl1_sleep.htm

or add the following to a bash script:

#!/bin/bash
sleep 20 && boinc & boincmgr
exit

Or use Zenity which is a graphical user interface (GUI) in the Gnome-desktop environment that returns the results to the console. Then you'd use something like this:

#!/bin/bash
zenity --question --timeout=30 --text="Starting programs in 30 sec." --title="Autostart"; STATUS=$?
if [ $STATUS = "1" ]; then
	echo ""
else
	boinc & boincmgr
fi

Explanation

#!/bin/bash 	Using the BASH-Shell to execute the commands. (but it doesn't has to be /bash)
--question 	Create a dialog with OK and Cancel button.
--timeout=30 	Wait 30 seconds for an input.
--text="" 	Text to show.
--title="" 	Title of window.
STATUS=$? 	Saves the return value in variable $STATUS.
if [ $STATUS = "1" ]; then 	If the return value is 1 (User hit "Cancel"), then ...
echo "" 	... print out nothing. (There has to be some command in the IF-branch.)
else 	... (Return value 0 = OK, 5 = TimeOut) ...
program1 & program2 & ... 	Starting programs at the same time (don't wait for program1 to be finished).
fi 	End of IF-branch. 
[code]

---