Maybe it was shortly after router reboot and it was doing clock adjusting during the time, since I am in GMT+1 timezone. The further content of file /tmp/log/ripe_sw_probe looks time-consistent. See attachment (.txt extension added. At 9:08 I rebooted the router). I uncommented those two lines in both scripts, now it does not show any messages in Log. When I run /etc/init.d/atlas restart, I get following log: Jan 9 10:21:25 turris atlas-sw-probe: Stopping atlas sw probe Jan 9 10:21:25 turris atlas-sw-probe: Kill all atlas processes Jan 9 10:21:25 turris atlas-sw-probe: Kill ssh tunnel Jan 9 10:23:37 turris telnetd[8892]: found atlas session Nothing else ATLAS related in Log. It looks the probe is now sending measurements results fine. Thanks for now, I will have further questions about the Probe network settings, will send another email shortly... Cheers Jiri ______________________________________________________________
Od: "Philip Homburg" <philip.homburg@ripe.net> Komu: atlas-sw-probes@ripe.net Datum: 09.01.2020 10:35 Předmět: Re: [atlas-sw-probes] SW Probe 1000069 issue
Hi,
On 2020/01/08 10:31 , ripe@brite.cz wrote:
Some outputs:
root@turris:~# /etc/init.d/atlas log 01/08/20 10:17:01 net-ok 01/08/20 09:17:01 reg-init 01/08/20 09:17:03 ctrl-init 01/08/20 09:17:04 keep-start 01/08/20 10:20:04 keep-found 01/08/20 10:20:04 net-ok 01/08/20 10:23:04 keep-found 01/08/20 10:23:04 net-ok 01/08/20 10:26:04 keep-found 01/08/20 10:26:04 net-ok
The weird thing here is the timestamps. I wonder why the clock seems to jump back for an hour for 3 of the entries.
Syslog: Jan 8 09:29:04 turris ATLAS[2929]: enough space free, no need to do anything Jan 8 09:29:04 turris ATLAS[2929]: buddyinfo: can't open '/proc/buddyinfo': No such file or directory Jan 8 09:29:04 turris ATLAS[2929]: buddyinfo: can't open '/proc/buddyinfo': No such file or directory Jan 8 09:29:04 turris ATLAS[2929]: condmv: not moving, destination '/usr/libexec/atlas-probe-scripts/data/out/v6addr.txt' exists Jan 8 09:29:04 turris ATLAS[2929]: condmv: not moving, destination '/usr/libexec/atlas-probe-scripts/data/out/simpleping' exists
This is mostly harmless. On hardware probes, those messages went to /dev/null. The last two should be gone in the next release of software probe code. I'll make a note to suppress the first one and make sure to disable code that generates the 2nd and 3rd on software probes.
When I look at our logs, I can see something weird that I don't quite understand.
The 'ATLAS ' and 'reginit.sh' scripts contain the lines (for the ATLAS script, lightly different for the reginit.sh script): #exec >/tmp/ATLAS.out 2>/tmp/ATLAS.err #set -x
Maybe you can uncomment those lines in both scripts, reboot and send me the logs files?
Philip