We noticed that the synchronization of the SpectroServer DB (to the 2nd server) during OLB does not work. The RCPD is not started, hence the 1st SpectroServer fails in syncing the SSDb to the 2nd. Having a closer looks the processd seems to crash, at least we see core dumps of the processd binary:
core.21418: ELF 64-bit LSB core file x86-64, version 1 (SYSV), SVR4-style, from '/opt/Spectrum/lib/SDPM/processd --start', real uid: 1678, effective uid: 1678, real gid: 1672, effective gid: 1672, execfn: '/opt/Spectrum/lib/SDPM/processd', platform: 'x86_64'
Workaround: restarting the processd (incl. all related programs, like the SpectroServer) seems to solve the issue for a while.
Release : 20.2, 21.2 (Linux)
Component : Spectrum Core / SpectroSERVER Fault Tolerance Online Backup
Code bug in processd that will crash starting in 21.2.8.
Set Processd_LCOPY = True to prevent processd from crashing as a workaround.
1. Open /opt/SPECTRUM/spectrum80.env file on secondary SS VM and then add "PROCESSD_LCOPY=TRUE" parameter and save the changes to it.
2. Stop & Start the processd service using systemctl
3. Now, start the OLB sync from OneClick and see whether the primary dB file is synchronized to secondary SS or not.
This can be tracked under DE537860 DE537125 and the final fix will be scheduled to be part of a future Spectrum release.