replication does not follow schedule

[Manny Vazquez]

Has anyone realized that does not matter the 'timing' you setup for the replication, it ends up falling out of its expected schedule to SYNC with round time?
This one shows next sync will be in both at 13:00 even that it is NOT supposed to happen at that time for the bottom one. 12:37 + 00:37 should be 13:14 (one minute before the other one)

But as you see here, it not only did the sync at 13:00 in both, it scheduled correctly for :15 and :37 ..



When :15 came around ..



Correct // but when :37 comes around, it resets again to :00

Look at this images, as you'll see, I have one to run */15 and the other */37 , even is both started at the same time, (which they did not) they should never (or very rarely) sync as it shows in this other screenshot


Does anyone know how the time is calculated in proxmox? it seems to be in a different ... zone

 

[wolfgang]

Hi,

the time calculation is the same as in systemd.time Calendar Event and this only allow a division of one hour.

 

[Manny Vazquez]

Hi,

the time calculation is the same as in systemd.time Calendar Event and this only allow a division of one hour.

So, basically you imply that there is no way to 'stager' the replications? Does not matter if you put 12 min, 15 or 17, eventually at the 00 they will all be at the same time? that seems to create a lot of work at the same time.

 

[wolfgang]

There is no parallel sync. Just one sync after another.
So if two syncs are scheduled at :00 randomly on will start and the other starts after the first are finished.

 

[Manny Vazquez]

There is no parallel sync. Just one sync after another.
So if two syncs are scheduled at :00 randomly on will start and the other starts after the first are finished.

Thanks..

I do not understand then why sometimes it tells me either timed out of could not get lock.. I am quite sure the next time around it syncs, but it is scary to have to check daily .

 

[wolfgang]

Please, can you send the error message?

 

[Manny Vazquez]

Sorry, I changed the replication times, the only real problem was the FIRST time, after that it is just seconds.