DrayTek UK Users' Community Forum
Help, Advice and Solutions from DrayTek Users
Vigor 3900 (1.0.6) Flaky and then Died
- freedom
- Topic Author
- Offline
- Junior Member
Less
More
- Posts: 14
- Thank you received: 0
21 Mar 2013 18:23 #75652
by freedom
Vigor 3900 (1.0.6) Flaky and then Died was created by freedom
I am currently in conversation with tech support regarding an issue I had but I just wanted to create a post here with updates and information that may be useful to others as I go along!
For a few weeks my 3900 has been a bit flaky. It would suddenly decide to stop responding to DHCP requests and then would stop working period. It would require a power cycle at this point to be made operational again.
Then suddenly on Wednesday it died completely and would no longer boot. I tried performing a factory reset and even tried a TFTP transfer with no joy.
I spoke to tech support who advised to try pushing it a couple of different firmwares via TFTP. Sadly this did not work.
We then were about to post the device back Thursday when tech support told me to try the serial console.
As you can see there was one error and I have no idea what it means:
Ignoring this we decided to see if we could use tftp to send some fresh fw, either 1.0.6.1RC3a build 2021 or 1.0.7 build 2005.
For a few weeks my 3900 has been a bit flaky. It would suddenly decide to stop responding to DHCP requests and then would stop working period. It would require a power cycle at this point to be made operational again.
Then suddenly on Wednesday it died completely and would no longer boot. I tried performing a factory reset and even tried a TFTP transfer with no joy.
I spoke to tech support who advised to try pushing it a couple of different firmwares via TFTP. Sadly this did not work.
We then were about to post the device back Thursday when tech support told me to try the serial console.
Code:
************************************
* DrayTek Vigor3900 Bootloader *
* Version: 0.1.3-20101109 *
************************************
Initializing NAND flash ...
NAND: 1024 MiB
Initializing Switch...
Switch init done
Press [ENTER] key within 5 sec. to download image... 0
Creating 1 MTD partitions on "nand0":
UBI: attaching mtd2 to ubi0
UBIFS error (pid 0): ubifs_recover_leb: LEB 228 scanning failed
DrayTek V3900 EEPROM v2.7
DDR Training.............................................................Done
As you can see there was one error and I have no idea what it means:
UBIFS error (pid 0): ubifs_recover_leb: LEB 228 scanning failed
Ignoring this we decided to see if we could use tftp to send some fresh fw, either 1.0.6.1RC3a build 2021 or 1.0.7 build 2005.
Please Log in or Create an account to join the conversation.
- freedom
- Topic Author
- Offline
- Junior Member
Less
More
- Posts: 14
- Thank you received: 0
21 Mar 2013 18:24 #75653
by freedom
Replied by freedom on topic Re: Vigor 3900 (1.0.6) Flaky and then Died
Code:
************************************
* DrayTek Vigor3900 Bootloader *
* Version: 0.1.3-20101109 *
************************************
Initializing NAND flash ...
NAND: 1024 MiB
Initializing Switch...
Switch init done
Press [ENTER] key within 5 sec. to download image... 4
Current LAN IP is 192.168.1.1
New IP: 192.168.7.1
Current Server IP is 192.168.1.10
New IP: 192.168.7.10
Current image is V3K91102.ALL
New Name: V3K9_V106.all
PHY 1000Mbit FD
Using comcerto_gemac0 device
TFTP from server 192.168.7.10; our IP address is 192.168.7.1
Filename 'V3K9_V106.all'.
Load address: 0x86000000
Loading: #################################
removed hashes to read easier
#############################
done
Bytes transferred = 24772656 (17a0030 hex)
Checksum pass, start writing image to flash...
Updating Firmware...
NAND erase: device 0 offset 0x200000, size 0x2000000
Erasing at 0x21e0000 -- 100% complete.
OK
NAND write: device 0 offset 0x200000, size 0x17a0000
Writing data at 0x199f800 -- 100% complete.
24772608 bytes written: OK
NAND erase: device 0 offset 0x3b000000, size 0x2000000
Skipping bad block at 0x3c420000
Erasing at 0x3cfe0000 -- 100% complete.
OK
NAND write: device 0 offset 0x3b000000, size 0x17a0000
Bad block at 0x3c420000 in erase block from 0x3c420000 will be skipped
Writing data at 0x3c7bf800 -- 100% complete.
24772608 bytes written: OK
Creating 1 MTD partitions on "nand0":
UBI: attaching mtd2 to ubi0
Loading file '/boot/uImage' to addr 0x827f8000 with size 2080452 (0x001fbec4)...
Done(0)
Verifying Checksum ... OK
Uncompressing Linux... done, booting the kernel.
login: : kernel.core_pattern = /tmp/%e.%p.%s.%t.core
: [BOOT] upgrade from 1.0.3.2 or below, update checksum
: [ready to login]
: 2011-01-01 00:00:52: (log.c.166) server started
Please Log in or Create an account to join the conversation.
- freedom
- Topic Author
- Offline
- Junior Member
Less
More
- Posts: 14
- Thank you received: 0
21 Mar 2013 18:25 #75654
by freedom
Replied by freedom on topic Re: Vigor 3900 (1.0.6) Flaky and then Died
We seemed to have flashed 1.0.7 successfully but several things stuck out to me:
It would appear that we have some bad flash memory but the system does seem to be dealing with it - could this have occurred and caused corruption and the problem in the first place?
In the instructions on how to flash through the serial console, the screenshots clearly show an updated EEPROM and Bootloader, 4.0 and 1.0.1-20130220 respectively. I raised this and was informed that the flash should update both and I concurred as the help file showed a screenshot of the terminal window showing it loading two files after the firmware flash:
/boot/ver_eeprom.txt and /boot/ver_uboot.txt
The screenshot then shows that the terminal writes these files and confirms the bootloader is up to date.
So we then decided to flash the 1.0.6.1RC3a build 2021 through the webgui that was now back up and running. This produced the following result.
It would appear that we have some bad flash memory but the system does seem to be dealing with it - could this have occurred and caused corruption and the problem in the first place?
In the instructions on how to flash through the serial console, the screenshots clearly show an updated EEPROM and Bootloader, 4.0 and 1.0.1-20130220 respectively. I raised this and was informed that the flash should update both and I concurred as the help file showed a screenshot of the terminal window showing it loading two files after the firmware flash:
/boot/ver_eeprom.txt and /boot/ver_uboot.txt
The screenshot then shows that the terminal writes these files and confirms the bootloader is up to date.
So we then decided to flash the 1.0.6.1RC3a build 2021 through the webgui that was now back up and running. This produced the following result.
Please Log in or Create an account to join the conversation.
- freedom
- Topic Author
- Offline
- Junior Member
Less
More
- Posts: 14
- Thank you received: 0
21 Mar 2013 18:25 #75655
by freedom
Replied by freedom on topic Re: Vigor 3900 (1.0.6) Flaky and then Died
Code:
The system is going down NOW !! Sending SIGTERM to all processes.
pluto[14589]: shutting down interface lo/lo ::1:500 pluto[14589]: shutting down interface lo/lo 127.0.0.1:4500 pluto[14589]: shutting down interface lo/lo 127.0.0.1:500
pluto[14589]: shutting down interface lan-lan1/lan-lan1 192.168.7.1:4500 pluto[14589]: shutting down interface lan-lan1/lan-lan1 192.168.7.1:500 miniupnpd[19095]: received signal 15, good-bye
rrdcollect[17930]: Signal received. Preparing to exit... pluto[14589]: ADNS process terminated by signal 15 pluto[14589]: leak: 2 * struct event in event_schedule() pluto[14589]: leak: myid string
pluto[14589]: leak: my FQDN
pluto[14589]: leak: 2 * struct event in event_schedule()
pluto[14589]: leak: ocspcerts path
pluto[14589]: leak: aacerts path
pluto[14589]: leak: certs path
pluto[14589]: leak: private path
pluto[14589]: leak: crls path
pluto[14589]: leak: cacert path
pluto[14589]: leak: acert path
Requesting system reboot.ocesses.
Restarting system.
DrayTek V3900 EEPROM v2.7
DDR Training.............................................................Done
************************************
* DrayTek Vigor3900 Bootloader *
* Version: 0.1.3-20101109 *
************************************
Initializing NAND flash ...
NAND: 1024 MiB
Initializing Switch...
Switch init done
Press [ENTER] key within 5 sec. to download image... 0
Upgrading firmware ...
NAND read: device 0 offset 0x3b000000, size 0x2000000
Bad block at 0x3c420000 in erase block from 0x3c420000 will be skipped
Reading data from 0x3d01f800 -- 100% complete.
33554432 bytes read: OK
NAND erase: device 0 offset 0x200000, size 0x2000000
Erasing at 0x21e0000 -- 100% complete.
OK
NAND write: device 0 offset 0x200000, size 0x2000000
Writing data at 0x21ff800 -- 100% complete.
33554432 bytes written: OK
Saving Environment to NAND...
Erasing Nand...Writing to Nand... done
Creating 1 MTD partitions on "nand0":
UBI: attaching mtd2 to ubi0
Loading file '/boot/uImage' to addr 0x827f8000 with size 2088464 (0x001fde10)...
Done(0)
Verifying Checksum ... OK
Uncompressing Linux... done, booting the kernel.
login: : kernel.core_pattern = /tmp/%e.%p.%s.%t.core
: [BOOT] upgrade from 1.0.3.2 or below, update checksum
: [ready to login]
Start conn_dect
: 2011-01-01 00:01:11: (log.c.166) server started
Please Log in or Create an account to join the conversation.
- freedom
- Topic Author
- Offline
- Junior Member
Less
More
- Posts: 14
- Thank you received: 0
21 Mar 2013 18:26 #75656
by freedom
As we can see it failed to update the EEPROM or Bootloader. Unfortunately it is now after 17:30 and support are closed, but I really would like to try and force an EEPROM / Bootloader update - does anyone know how?
Note: Your message contains 13205 characters. The maximum number of allowed characters is 3000. <-- Annoying but I understand!
Replied by freedom on topic Re: Vigor 3900 (1.0.6) Flaky and then Died
Code:
The system is going down NOW !!
Sending SIGTERM to all processes.
pluto[15126]: shutting down interface lo/lo ::1:500
pluto[15126]: shutting down interface lo/lo 127.0.0.1:4500
pluto[15126]: shutting down interface lo/lo 127.0.0.1:500
pluto[15126]: shutting down interface lan-lan1/lan-lan1 192.168.7.1:4500
pluto[15126]: shutting down interface lan-lan1/lan-lan1 192.168.7.1:500
miniupnpd[19512]: received signal 15, good-bye
rrdcollect[18376]: Signal received. Preparing to exit...
Sending SIGKILL to all processes.
Requesting system reboot.
Restarting system.
DrayTek V3900 EEPROM v2.7
DDR Training.............................................................Done
************************************
* DrayTek Vigor3900 Bootloader *
* Version: 0.1.3-20101109 *
************************************
Initializing NAND flash ...
NAND: 1024 MiB
Initializing Switch...
Switch init done
Press [ENTER] key within 5 sec. to download image... 0
Creating 1 MTD partitions on "nand0":
UBI: attaching mtd2 to ubi0
Loading file '/boot/uImage' to addr 0x827f8000 with size 2088464 (0x001fde10)...
Done(0)
Verifying Checksum ... OK
Uncompressing Linux... done, booting the kernel.
login: : kernel.core_pattern = /tmp/%e.%p.%s.%t.core
: [BOOT] upgrade from 1.0.3.2 or below, update checksum
: [ready to login]
As we can see it failed to update the EEPROM or Bootloader. Unfortunately it is now after 17:30 and support are closed, but I really would like to try and force an EEPROM / Bootloader update - does anyone know how?
Note: Your message contains 13205 characters. The maximum number of allowed characters is 3000. <-- Annoying but I understand!
Please Log in or Create an account to join the conversation.
- sicon
- Offline
- Contributor
Less
More
- Posts: 642
- Thank you received: 0
22 Mar 2013 12:17 #75668
by sicon
Replied by sicon on topic Re: Vigor 3900 (1.0.6) Flaky and then Died
How did you get on?
Please Log in or Create an account to join the conversation.
Moderators: Chris, Sami
Copyright © 2024 DrayTek