[SOLVED] pve replication - invalid json

[rafafell]

Hi,

one node A down, the vms were being replicated on node B. However, when viewing the replication through the graphical interface, the following message appears:

invalid json data in '/var/lib/pve-manager/pve-replication-state.json' (500)

the json file is corrupted:

pst0 ^ D ^ K ^ H12345678 ^ D ^ H ^ H ^ H ^ C ^ B ^ @ ^ @ ^ @ ^ H� ^ H ^ @ ^ @ ^ @ maxDelay ^ Fz ^ VF _ ^ @ ^ @ ^ @ ^ @ ^ K ^ @ ^ @ ^ @ nextRunDate ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ ^ @ $

I moved the configuration from node A to node B. Apparently everything works fine.

I have the following doubts:

(1) can I recover the pve-replication-state.json file?
(2) Is it a good practice to have a backup of the pve-replication-state.json file?
(3) when Node A returns, will I be able to use the vm that was in it instead of the replica that is now active in node B?


update:

I made a copy of the file pve-replication-state.json and deleted the information from json leaving only {} and the graphical replication interface message reappeared (with apparently correct information), is this normal?


best,

 

[fiona]

Hi,
glad you were able to solve it yourself. Could you please mark the thread as [SOLVED] so others can find solutions more easily.

(1) can I recover the pve-replication-state.json file?
(2) Is it a good practice to have a backup of the pve-replication-state.json file?

update:

I made a copy of the file pve-replication-state.json and deleted the information from json leaving only {} and the graphical replication interface message reappeared (with apparently correct information), is this normal?

Yes, the information for the replication state can mostly be recovered automatically, and we also detect if the configuration has moved to another node. It saves things like:

• when was the last sync?
• did the last sync succeed?
• which storages were involved?
• etc.

(3) when Node A returns, will I be able to use the vm that was in it instead of the replica that is now active in node B?

You will lose the writes on node A since the last sync once you replicate back from B to A.