same here. Samba is probably compiled without IPC support. This is problematic for share folder browsing, but not for connecting and using your share.
Hi,
I upgraded firmware to 1.8.1.7-3 and I started to get very awfull messages in /var/log/smb - seems like shared memory is somewhat corrupt (?).
The smbd process seems to span more and more children processes.
The smb.conf was not changed from that which worked in 1.8.1.7-2a and previous.
Access to shared folder is working well, despite of errors, but it cease to function after some time.
Can anybody tell me how to correct this? Or samba itself is broken?
/etc/smb.conf
/var/log/smb:Code:[global] netbios name = FW server string = workgroup = HOME local master = yes prefered master = yes os level = 255 guest account = nobody security = share browseable = yes log level = 1 max log size = 100 encrypt passwords = no dns proxy = no wins support = yes preserve case = yes short preserve case = yes [share] comment = Shared public folder path = /mnt/share writeable = yes guest ok = yes force user = nobody
Code:WARNING: clearing global IPC lock set by dead process 15406 ERROR shm_open : can't IPC_SET. Error was Bad address 2004/11/21 16:28:35 pc102 (192.168.11.102) connect to service share as user nobody (uid=99,gid=99) (pid 15807) 2004/11/21 16:28:35 pc102 (192.168.11.102) couldn't find service shar 2004/11/21 16:28:35 pc102 (192.168.11.102) couldn't find service shar 2004/11/21 16:29:09 pc102 (192.168.11.102) closed connection to service share ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! 2004/11/21 16:29:13 pc102 (192.168.11.102) connect to service share as user nobody (uid=99,gid=99) (pid 15809) 2004/11/21 16:29:20 pc102 (192.168.11.102) couldn't find service shar 2004/11/21 16:29:30 pc102 (192.168.11.102) closed connection to service share ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! 2004/11/21 16:30:57 pc102 (192.168.11.102) connect to service share as user nobody (uid=99,gid=99) (pid 15817) 2004/11/21 16:30:57 pc102 (192.168.11.102) couldn't find service shar 2004/11/21 17:15:36 pc102 (192.168.11.102) closed connection to service share ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! =============================================================== INTERNAL ERROR: Signal 11 in pid 16545 (1.9.18p10) Please read the file BUGS.txt in the distribution =============================================================== Last message was pSETDIR =============================================================== Dumping core in /var/log/corefiles write_socket: Error writing 4 bytes to socket 5: ERRNO = Connection reset by peer Error writing 4 bytes to client. -1. Exiting WARNING: clearing global IPC lock set by dead process 16545 ERROR shm_open : can't IPC_SET. Error was Bad address ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! write_socket: Error writing 4 bytes to socket 5: ERRNO = Connection reset by peer Error writing 4 bytes to client. -1. Exiting
same here. Samba is probably compiled without IPC support. This is problematic for share folder browsing, but not for connecting and using your share.
1.8.1.7-3 uses "newer" samba, and it has some problems. I've not time to check this yet.
Same problems here too. I suffered a huge performance loss, because the time needed before the datatransfers actually starts is very long. Data transfer itself is pretty fast, I reached 800 kByte/s while reading. Thats 200 more then with 1.7.x.x series firmware.
My smb log:
I get the messages above very often, so below is the last bit of the logCode:Unknown message type 160! 2004/11/21 19:09:19 unknown command type (SMBunknown): cnum=105 type=160 (0xA0) Unknown message type 160! 2004/11/21 19:09:19 unknown command type (SMBunknown): cnum=105 type=160 (0xA0) Unknown message type 160!Code:Unknown message type 160! 2004/11/21 19:10:23 unknown command type (SMBunknown): cnum=105 type=160 (0xA0) 2004/11/21 21:05:05 tester (192.168.1.10) closed connection to service share 2004/11/21 23:36:05 travelmate (192.168.1.2) closed connection to service share$ ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! 2004/11/21 23:37:17 travelmate (192.168.1.2) connect to service share$ as user admin (uid=0,gid=99) (pid 370) 2004/11/21 23:44:27 travelmate (192.168.1.2) closed connection to service share$ ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! ERROR shm_validate_header : bad magic ERROR shm_open : corrupt IPC area - remove it! 2004/11/22 21:36:31 travelmate (192.168.1.2) connect to service share$ as user admin (uid=0,gid=99) (pid 1475)
I just discovered a 'corefiles' directory in the '/var/log' directory. It contains a file called 'core' which is 417792 bytes big. I have no idea what it is. I can find on Google many ways how to clean them, but nowhere what they are. They seem to me like core dumps when a program crashes or something. Am I right? Are these files nessesary or can I safely delete them, because they consume lots of ram...?
A core file is an in-memory dump of the complete process image at the time that process was forcefully terminated by am unhandled, unmasked signal (most often "signal" would equate what often is referred to as "crash" ;-)).Originally Posted by Styno
The only common use I am aware of would be "after-the-fact" diagnostic - if you were to send a given (unstripped (of debug symbol information) binary to somebody who has access to source code and a development environment, the recipient of this core file could then be able to load the core file into GDB (the GNU debugger) and look at, say, stack contents, the program counter, or register contents at the time the unhandled signal. Obviously this aids in debugging, but it comes close to reading tea leafs of some kind.
These core files can be safely deleted, any time.
Yesterday I tried to flash version 1.8.1.7-3 but encountered many problems. One of the problems was that the admin / admin account did not work anymore; I had to add the admin user. After a full reset (holding the reset button for more than 15 seconds) I flashed it again, now I could log in into the web administration but the telnet console did not work anymore (admin admin). I had to add user admin again, then I was able to log in. However in the CLI i could not find the /mnt/usbfs anymore, while ftp did work to the usb stick. Did I missed something?
Yes, I think you missed somethingOriginally Posted by bramfm
I guess you upgraded from a 1.7.x.x version firmware to this one?! After flashing to the new 1.8.x.x. series firmware always reset to defaults. After that the admin password should just work fine.
The 1.8.x.x. series firmware differ quite a bit from the older ones: The USB storage device is mounted on /tmp/harddisk for example. You can mount it the old fashioned way too, but you'll have to configure that yourself (Oleg has described how-to do this somewhere else).
Originally Posted by Styno
I think I did not look futher than the length of my nose (as we say in Dutch). Thanks!
Hi Oleg.Originally Posted by Oleg
If possible, can you set this back to a fixed value, say 5 retrys and make a new flash variable for this to change, since some isp´s kill the acces when slamed with requests. After 10 atempts with wrong password for example, my provider bans the acces for 24h and unlocking before this timeout is a hassle... Or is this not affected with this retry value ?
And, can you please make the uptime display in the status page back to the pre-1.7.x.x look ? (hours and minutes, not secounds only) .
Greets
My Stuff: WL-500g, Mapower H31x 10GB HD, Philips Webcam Vesta PRO, TerraTec Webcam PRO, USB Hub
I´m using Oleg´s 1.8.1.7 CR3.
Since upgrading my WL500g crashes up to every five minutes!
Anyone with the same problem?
Turn off you wl500g for 10 minutes and then turn it on.
By turning off my WL500g for 10 minutes one time this problem is solved? why?
Well, thank you for fast answaering oleg ;-)
Most likely it has detected not the whole RAM after upgrade. This is a feature of broadcom memory size detection algorithm.
I've flashed my asus with this version about a week ago.
At first I thought I suffer the wan port die thingy as it randomly drops internet connection. So I put the speed to a fixed rate.
However, since I use automatic IP (modem is in bridge mode and ISP gives out dhcp adres) I see these error messages:
Dec 24 18:00:18 ntp client: time is synchronized to ntp.demon.nl
Dec 24 19:00:05 dhcp client: deconfig: lease is lost
Dec 24 19:00:06 dhcp client: bound IP : 82.161.x.yy from 82.161.6.1
Dec 24 20:00:05 dhcp client: deconfig: lease is lost
Dec 24 20:00:06 dhcp client: bound IP : 82.161.x.yy from 82.161.6.1
Dec 24 20:00:18 ntp client: time is synchronized to ntp.demon.nl
Dec 24 21:00:11 dhcp client: deconfig: lease is lost
Dec 24 21:00:12 dhcp client: bound IP : 82.161.x.yy from 82.161.6.1
Dec 24 22:00:10 dhcp client: deconfig: lease is lost
Dec 24 22:00:11 dhcp client: bound IP : 82.161.x.yy from 82.161.6.1
Dec 24 22:00:24 ntp client: time is synchronized to ntp.demon.nl
Dec 24 23:00:15 dhcp client: deconfig: lease is lost
Dec 24 23:00:16 dhcp client: bound IP : 82.161.x.yy from 82.161.6.1
So it seems it has a problem with renewing the dhcp lease, so roughly every now and then it drops the wan connection.
Anyone an idea how to fix this?
I will try another firmware tomorrow to see if that helps.