Hi again!
I've managed to reproduce this by hand, by trying to clone the same repos as root, on the same host. It seems that the use of "--depth=1" makes the command fail. When --depth is not used, the clone works correctly:
root@eltf-qemu01:~/tmp_lava# git clone git://git.enea.se/eltf/lava.git /var/lib/lava/dispatcher/tmp/384/lava-overlay-SXDmxh/lava-384/0/tests/0_healthcheck-runner --depth=1 Cloning into '/var/lib/lava/dispatcher/tmp/384/lava-overlay-SXDmxh/lava-384/0/tests/0_healthcheck-runner'... fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed
root@eltf-qemu01:~/tmp_lava# git clone git://git.enea.se/eltf/lava.git /var/lib/lava/dispatcher/tmp/384/lava-overlay-SXDmxh/lava-384/0/tests/0_healthcheck-runner Cloning into '/var/lib/lava/dispatcher/tmp/384/lava-overlay-SXDmxh/lava-384/0/tests/0_healthcheck-runner'... remote: Counting objects: 1253, done. remote: Compressing objects: 100% (1241/1241), done. remote: Total 1253 (delta 675), reused 0 (delta 0) Receiving objects: 100% (1253/1253), 239.59 KiB | 0 bytes/s, done. Resolving deltas: 100% (675/675), done. Checking connectivity... done.
Is there a way to avoid the use of the --depth parameter (i.e. a setting), or should I just make the changes in the LAVA code?
Thanks again! -- /Dragoş
On 2017-11-06 11:30:35+02:00 Dragos Iorgulescu wrote:
Hi!
Just upgraded from 2017.6 to 2017.10-1. Tried re-running some jobs that worked fine just before the upgrade and noticed that git checkouts can no longer pe performed.
The encountered error is: fatal: The remote end hung up unexpectedly\nfatal: early EOF\nfatal: index-pack failed\n
Tried checking out manualy on the host, using the same repo. It works.
How could I proceed investigating this? Could this be related to some lava user configuration that got reset (I don't know if git-related configs are associated with LAVA)?
Thanks in advance (again)! P.S: I have attached the plain job log -- /Dragoş