Hello,
In an attempt to resolve the issues with segmentation fault I mentined in another thread (MGM segmentation fault), I upgraded to 4.8.39 and recreated the QuarkDB dirs. The QuarkDB cluster starts up ok and a chorum is formed.
However, when I start eos@mgm on the single MGM node I have, the process starts (and there is a link to the leader QuarkDB node) but I can’t see a message in /var/log/eos/mgm/xrdlog.mgm reporting that the master lease has been aquired.
I do see the following in the /var/log/messages, I have tried quite a few things without much sucess. If you could give me a clue that would be really great. Many thanks, George
Mar 22 18:10:35 host-172-16-103-63 systemd: Starting EOS mgm…
Mar 22 18:10:35 host-172-16-103-63 systemd: Started EOS mgm.
Mar 22 18:10:35 host-172-16-103-63 eos_start.sh: Using xrootd binary: /opt/eos/xrootd/bin/xrootd
Mar 22 18:10:35 host-172-16-103-63 eos_start.sh: Register objects provided by NsQuarkdbPlugin …
Mar 22 18:16:16 host-172-16-103-63 kernel: xrootd[11929]: segfault at a0 ip 00007f4ee9464129 sp 00007f4eedbe7b20 error 4 in libEosCommon.so.4.8.39[7f4ee92d1000+41c000]
Mar 22 18:16:16 host-172-16-103-63 systemd: eos@mgm.service: main process exited, code=killed, status=11/SEGV
Mar 22 18:16:16 host-172-16-103-63 systemd: Unit eos@mgm.service entered failed state.
Mar 22 18:16:16 host-172-16-103-63 systemd: eos@mgm.service failed.
Mar 22 18:16:21 host-172-16-103-63 systemd: eos@mgm.service holdoff time over, scheduling restart.
Mar 22 18:16:21 host-172-16-103-63 systemd: Stopped EOS mgm.
Mar 22 18:16:21 host-172-16-103-63 systemd: Starting EOS mgm…