https://turingpi.com logo
#│forum
Chaining flashing of CM4 fails after 1st node
# │forum
n

ncasaux_96421

12/17/2023, 11:58 PM
Hello, I'm using a local
tpi
to attempt to flash, 1 by 1, all the 4 nodes of the Turing Pi 2. It always fails quicky after starting to upload image on node 2. Here my console output:
Copy code
for i in 1 2 3 4
do
        tpi --host 192.168.1.100 --user root --password turing flash --image-path /tmp/raspios.img -n $i
done
request flashing of raspios.img to node 1
started transfer of 2.55 GiB..
⠤ [00:09:09] [#######################################################################################################################################################################################################>] 2.55 GiB/2.55 GiB (0.0s)⠁ Verifying checksum...
Done
request flashing of raspios.img to node 2
started transfer of 2.55 GiB..
⠖ [00:00:24] [###>------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------] 48.94 MiB/2.55 GiB (1259.0s)error sending request for url (https://192.168.1.100/api/bmc/upload/3179651044): connection error: Connection reset by peer (os error 104)
request flashing of raspios.img to node 3
error sending request for url (https://192.168.1.100/api/bmc/authenticate): error trying to connect: Connection reset by peer (os error 104)
request flashing of raspios.img to node 4
error sending request for url (https://192.168.1.100/api/bmc/authenticate): error trying to connect: tcp connect error: Connection refused (os error 111)
On the BMC side, I have those logs. It ends with a
Dec 17 23:47:55 turingpi kern.info kernel: [ 1069.347412] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),task=bmcd,pid=1142,uid=0
Is there something I could do to achieve this, or is it a bug ? https://cdn.discordapp.com/attachments/1186095026573410335/1186095027349364807/message.txt?ex=65920013&is=657f8b13&hm=7cafd95e307d2991dd79d90a91032c092bb54d8ea0bd1b56f5be54a9abeb56c9&
@User , do you think this is a bug ? Any idea how I could find a workaround ? Maybe include a reboot of BMC after each install ? 🥲
s

svenrademakers

12/19/2023, 5:36 PM
well thats interesting, it seems the BMC ran out of memory...
can you reproduce this consistently?
what happens if you put a sleep of a min or so between the flash commands?
i created a ticket on github for visibility, see https://github.com/turing-machines/BMC-Firmware/issues/162
n

ncasaux_96421

12/19/2023, 5:46 PM
Thanks @svenrademakers , I'll try with a 60s sleep and let you know!
Just tested, it does exactly the same thing. I added a comment on the Github issue directly.
4 Views