PDA

View Full Version : Linux client not uploading



TheOtherPhil
05-24-2002, 01:51 AM
I have tried both the gcc and Intel compiled client on my Gentoo linux system. I am using the correct handle and work is downloaded OK but it never gets uploaded. Once it has ran 25000, it will stop and report that it can't reach the network. This system is on the same switch as my windows boxes and I have full internet functionality. My Windows machines are running fine.

Any ideas?

TheOtherPhil
05-24-2002, 02:32 AM
Make that 30,000 before stopping :help:

Brian the Fist
05-24-2002, 12:44 PM
Send your error.log file to us at trades@mshri.on.ca and are you behind a proxy sevrer/firewall?

TheOtherPhil
05-24-2002, 01:53 PM
Thanks Howard, here are the contents of my error logs if anybody else has an idea:



CPU0

========================[ May 23, 2002 7:59 PM ]========================
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status

========================[ May 24, 2002 5:36 AM ]========================
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status

========================[ May 24, 2002 5:37 AM ]========================
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 4991} Too many files buffered



CPU1

========================[ May 23, 2002 8:00 PM ]========================
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status

========================[ May 24, 2002 5:37 AM ]========================
ERROR: [000.000] {foldtrajlite.c, line 1129} Unable to check server status
!!! Contact between residues LEU and THR too close !!!

Basically, the client will connect to the server initially to get work, but will not contact it again. It will run all 30,000 units and then just stop and complain about no network or invalid handle (the handle is correct). I have tried both the gcc and intel compiled client. The machine is sat on the same switch as my Windows boxes and the network settings are the same. The 6 windows clients work fine. Internet access is working on the linux box. I use a Netgear RT314 router which shares my 1Mbit cable connection to my switch.

MAD-ness
05-24-2002, 06:49 PM
Step 1:

Edit the "fodlit.bat" batch file in the distribfold\ directory. You will see something like this at the start of the file:


@echo off
:start
.\foldtrajlite -f protein -n native

The third line contains 'parameters' or 'arguments' (not sure what the proper term would be) for running the client .exe (foldtrajlite).

You want to add a "-df" 'switch' to this line so that it looks like this:


.\foldtrajlite -f protein -n native -df

This will enable much larger buffering, which will allow you avoid the error message and halting of the client after 6 sets of structures are completed (30,000 proteins). The -df switch will allow your client to keep on crunching, even if it isn't able to contact the server to upload results. Also, if you have a larger number of structures to return you NEED the -df switch in order to properly upload the results (otherwise it can 'lose' some of them). With the -df switch enabled you don't have to worry about losing results like this.

You seem to have assumed that the client is connecting to "get work" but this client does NOT download work units, once you install the client you do not have to connect to the DF server before you can begin crunching. Just an FYI I guess.

I do not know why you were unable to connect to the server, but eventually did (when you uploaded results), so someone else can hopefully help there.

But the client halting at 30,000 results and then "losing" results when uploading these 30k structures/proteins/results can be avoided in the future by setting the "-df" switch in the 'Foldit.bat' file.

Good luck.

TheOtherPhil
05-25-2002, 01:19 AM
Thanks for the info, but I am still unable to send any results. I know my network connection is fine because I downloaded the client with the wget command and my Windows machines are uploading fine. I am going to install RH7.3 soon anyway, so this may fix my problem.

Ragnarog
05-25-2002, 02:53 PM
The DF servers themselves are down right now...so..noone can upload I think...

Rag

Terminator
05-25-2002, 02:58 PM
Lo Rag ! How you doing m8 ? - He was having problems before the server borked ;) I managed to get the ICC version running on RH7.3 (the new 16th May client) and was uploading OK before the server started refusing to accept uploads :)

Ragnarog
05-25-2002, 03:02 PM
I should have checked the dates perhaps? :D

Everything's fine here Termy :)

Am storming up the DF charts at 3DHQ ;)
Remember the Wacky Races on genome next month! :D

Rag

Brian the Fist
05-25-2002, 03:19 PM
Server problem is fixed, I believe

RipItUp
05-26-2002, 04:25 AM
One of my machines keeps getting a Windows XP version of a gpf where it wants to send an email to William Gates.

I aslo had the server issue of course mentioned above.

So now I have several BZ files that will not send in, are these unrecoverable ? The number on the .val.bz2 file, is that the size or just a unique id ?

Sorry for the dumb ass question but I haven't got time to do much searching as I have nappies to change nowadays !

Many thanks for any help and thanks for the DF tip above, I read it in the readme but didn't realise it also helped send all the files rather than just let you keep going offline.

Regards

Andy

MAD-ness
05-26-2002, 10:53 PM
The number at the start of the bz2 files is, I think, your handle.

Also, you can collect orphan files....a thread around here and a faq on www.teamstirfry.com contain directions for this.

Note, you can/will have more .val files than .log files....the # of .log files is the number of results sets.

.val files don't actually have results.

TheOtherPhil
05-27-2002, 03:33 AM
Thanks for the help everybody. I was unable to get my Gentoo Linux box to upload and couldn't see anything immediately obvious. I decided to install a copy of Mandrake 8.2 that I had to hand and I still get the same problem.....nothing gets uploaded and the client thinks there is no net connection. The Linux box is configured in exactly the same way network wise as my other Windows machines and is also connected to the same switch/router. I have full internet capability (web, mail, ftp, ssh etc are working) so I can't see the problem.


As a work around, I am harvesting the work via ftp to my windows box and uploading it from there. It is not exactly an ideal solution, but at least it is working.

fimmi
05-27-2002, 09:09 AM
i have the exact same problem! :help: and i know of others who have it too. i'm running mandrake :bang: :bang: :bang:

TheOtherPhil
05-27-2002, 06:46 PM
Originally posted by fimmi
i have the exact same problem! :help: and i know of others who have it too. i'm running mandrake :bang: :bang: :bang:


Yup, just installed Mandrake.....I was running Gentoo previously and both distro's have the same fault. :help:

TheOtherPhil
08-07-2002, 12:14 PM
I am still having this problem with my Linux machines....I have tried an install of SUSE8 Pro with the same results....I cannot upload results and the client refuses to contact home.

What ports does the client need open for general use? It isn't a problem with my firewall as my Windows machines work fine so I am assuming that Linux is setting up some sort of default firewall?

Argh! :bang:

dnar
08-07-2002, 01:03 PM
Do you have a default route setup?

What does /sbin/route produce?

TheOtherPhil
08-07-2002, 01:40 PM
Originally posted by dnar
Do you have a default route setup?

What does /sbin/route produce?


Hi Wayne, here's what I get:



Tiger3:/home/phil # /sbin/route
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.0.0 * 255.255.255.0 U 0 0 0 eth0
default 192.168.0.1 0.0.0.0 UG 0 0 0 eth0

I have tried manually setting my network info and I currently have it running DHCP with my router (192.168.0.1) acting as the DHCP server. The router is set with my ISP DNS server info. I have full internet usage (ftp/web etc) but just cannot get anything from the DF client under *any* Linux distro. :help:

dnar
08-08-2002, 04:41 AM
Looks fine. It you run the DF client while tailing your logs you will see if you have an IPTABLE block on the port.

su <passwd>
tail -f /var/log/messages

TheOtherPhil
08-08-2002, 09:30 AM
Originally posted by dnar
Looks fine. It you run the DF client while tailing your logs you will see if you have an IPTABLE block on the port.

su <passwd>
tail -f /var/log/messages

The only thing I get when I start/stop the client is
Aug 8 14:25:17 Tiger3 modprobe: modprobe: Can't locate module char-major-10-134

I don't think that this is anything to do with my problem though.

dnar
08-08-2002, 01:21 PM
According to my /dev system, 10,134 is the apm_bios device. What is the client doing with power management? :eek:

dnar
08-08-2002, 01:23 PM
You can resolve domain names from the Linux machine yes?

TheOtherPhil
08-08-2002, 01:28 PM
Originally posted by dnar
You can resolve domain names from the Linux machine yes?


phil@Tiger3:~> ping www.distributedfolding.org
PING distributedfolding.org (206.248.62.6) from 192.168.0.40 : 56(84) bytes of data.
64 bytes from 206.248.62.6: icmp_seq=1 ttl=239 time=118 ms
64 bytes from 206.248.62.6: icmp_seq=2 ttl=239 time=112 ms
64 bytes from 206.248.62.6: icmp_seq=3 ttl=239 time=113 ms

--- distributedfolding.org ping statistics ---
3 packets transmitted, 3 received, 0% loss, time 2015ms
rtt min/avg/max/mdev = 112.608/114.901/118.463/2.567 ms


Yep, every other service works as expected....web, ftp, ssh etc.

dnar
08-08-2002, 01:41 PM
Shoot man, I dont know... Maybe I'll just have to jump on a jet plane and come take a look. :swear: :bang: :swear:

TheOtherPhil
08-08-2002, 01:43 PM
Originally posted by dnar
Shoot man, I dont know... Maybe I'll just have to jump on a jet plane and come take a look. :swear: :bang: :swear:


Cool....I'll stick some tinnies in the fridge :) Thanks for trying though :cool:

dnar
08-08-2002, 02:27 PM
You could try deleting all IPTABLES rulesets.

/sbin/iptables -X
/sbin/iptables -F

If it's not a firewall issue, you have a route setup and you can resolve IP's, then I am at a loss... The beer in the fridge is tempting, it's only what, 20,000kms?

TheOtherPhil
08-09-2002, 01:03 PM
Hmmm, IPTables isn't even installed. This is sooo annoying now. I have exactly the same problem on Red Hat 7.2/7.3, Gentoo, Mandrake 8.1/8.2 and Suse 8 using either client. My Windows machines on the same network run great :bang:

dnar
08-09-2002, 01:19 PM
Well if IPtables aint installed (and I doubt your newer distros will install IPchains) and your having this problem with ALL Linux boxen, I would have to say that is weird.

My advise, install Windows, Linux sux. :D

Brian the Fist
08-10-2002, 11:42 AM
Phil,

I was not aware until just now that it was ALL your Linux machines including RedHat which were non-functional. That is indeed very unusual. If you wish me to try to assist, try using a port sniffer to get an ASCII dump of the packets going between one of your boxes and www.distributedfolding.org on port 80 when you start up the program. Also send me your proxy.cfg file if you have one in the directory of the program. This'll give me a better idea of what could be wrong. Thanks

TheOtherPhil
08-12-2002, 04:08 PM
Thanks Howard...I'll have a root about with a packet sniffer tomorrow sometime.

DB7654321
09-07-2002, 01:54 AM
I can confirm this problem on a new distfolding install. I'm running Debian Linux (3.0, aka Woody), and all other services are working properly. When I run tcpdump on my firewall box (Another Debian install) to catch all traffic through ppp0 (My dialup connection) and start the DF client, no packets are reported. The client is erroring without attempting a connection. My handle is correct (I copy and pasted it from the confirmation email). If you need additional info, feel free to email me.

DB7654321
09-07-2002, 02:53 AM
Nevermind -- Evidently, I didn't consider the effect of my transparent proxy. :) I have seen similar results for applications that send a HTTP request in the form of:

GET / HTTP/1.0

instead of

GET http://servername.com/ HTTP/1.0

...Which my proxy, Squid, prefers.