Last index records lost on server reboot

[d47081]

Hello,

I have one service where collecting index in manticore (instead of mysql) but found after server hard reboot, last records not available anymore, even crawled.

What could it be? Maybe I must setup some tmp buffer?

Just want to use manticore as the SQL database replacement, but magic - it hides last records after system restart.

p.s. I'm using crontab command @reboot rm /var/lib/manticore/binlog/*
because have issue with manticore start #1716 - maybe the problem in that

[sanikolaev]

Perhaps another binlog_flush strategy can be helpful https://manual.manticoresearch.com/Logging/Binary_logging#Binary-flushing-strategies

[d47081]

Thanks for info, I've disabled bin log by binlog_path = # empty for test.

Suppose the problem not in binlog timing / strategy, because record exist about few hours, then lost on system restart.
Maybe that is really depends of hard removing this file by rm - let me test with disabled one (because when enabled I even can't start the manticore server at all)

[d47081]

I have just understand that binlog crashes on the 'writing' second where default strategy is:

2 - Flush every transaction and sync every second. This offers good performance, and every committed transaction is guaranteed to be saved in the case of a server crash. However, up to 1 second's worth of committed transactions may be lost in the event of an OS/hardware crash.

Then, probably, I can't start manticore because binlog get corrupted. Of course rm is just forgotten in crontab tmp solution to launch the server with corrupted binlog.

In other words, instead of disable binlog at all, better to change it strategy to binlog_flush = 1 in my case (where I'm running the web crawler that works 24/7)

Thanks much!