Following a ptest failure, retrieving artifacts failed and actually stuck the rest of the build: https://autobuilder.yoctoproject.org/typhoon/#/builders/81/builds/6806/steps/13/logs/stdio log.do_testimage.311128 ended up being 4.3G DEBUG: [Running]$ ssh -l root -o ServerAliveCountMax=2 -o ServerAliveInterval=30 -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o LogLevel=ERROR 192.168.7.20 export PATH=/usr/sbin:/sbin:/usr/bin:/bin; tar zcf - /var/log /etc DEBUG: Waiting for process output: time: 1721087387.1578567, endtime: 1721087687.1568744 tar: removing leading '/' from member names DEBUG: Partial data from SSH call: <garbage>
There are some artifact collection code changes in master-next that Alex should be aware of.
I discussed a bit the issue with Alexandre, and despite "hung SSH session" likely being the suspect, I do not see how it could happen, considering the parameters passed to ssh commandline in oeqa scripts (connection keepalive and timeout parameters). I did not manage neither to reproduce locally the issue. As Randy mentioned, there has been some changes in the mean time which have reduced the number of ssh connections involved in artifacts retrieval (https://git.yoctoproject.org/poky/commit/?id=6e80b2ab660e60135b52bd7358f0ce7612226f06). Has the issue been observed recently (ie after those changes) ? If not, I suggest to close the issue.
Not seen again, closing for now