 |
The output_formatter has been added to a able to represent the output of bconsole commands in different formats. Normally you see the plain text output in the bconsole. With ".api json" you can switch to the json output. This error message occurs, when the OutputFormatter fail to send/print/write a message. Normally this is triggered by bconsole commands. Given the time of your message: is there a job with runscripts/admin job running? If yes, please send me this part of the configuration. Maybe there is some bug. Anyhow, it should not cause any harm, only the message will not be displayed. The error message states: "Maybe result message to long?" because there is a known issue with the underlying TCP functions, that prevent sending long (>= 64000 bytes?) messages. However, as you message is only 231 bytes, that is not the cause of your issue. |
|
|
I am very grateful for the answer. Was not enough information about "output_formatter", and so thank you - it became clear for what and why, and how going on. For my part, I want to apologize because we had a damaged (corrupted) database, (table "file"), and after restored database, errors ("Failed to send message")no longer appear in the log. -- "Given the time of your message: is there a job with runscripts/admin job running?" Yes you are right. But I am now not sure, because there were problems with the database... We use the following: *** JobDefs { Name = "postgres-dump" Type = "Backup" Level = Full Priority = 10 Messages = "bareos-dir-messages-standard" Storage = "disk-bkx1" Schedule = "postgres-dump" Pool = "postgres-dumps" Fileset = "postgres-dump-bpipe" Accurate = yes Allow Duplicate Jobs = no Cancel Queued Duplicates = yes Client Run Before Job = "/etc/bareos/scripts/get_postgres_mode.sh" } FileSet { Name = "postgres-dump-bpipe" Include { Options { Signature = SHA1 } Plugin = "\\|/etc/bareos/scripts/postgres-bpipe" } } *** |
|
|
Problem seems was on our side. joergs, thanks again. |
- Anonymous
