Tomas Barton

@oxzi It happens quite frequently (e.g. daily), the database runs on a dedicated physical server. There's basically no load at all until queries from both `icingadb` cause a deadlock. icingadb...

here's full log of another drive in similar condition: ``` smartctl -a /dev/sdb smartctl 6.6 2016-05-31 r4324 [x86_64-linux-4.9.0-16-amd64] (local build) Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org === START...

default check is failing due to `Percent_Lifetime_Used` ``` ./check_smart.pl -i ata -d /dev/sdb CRITICAL: Drive Crucial_CT250MX200SSD1 S/N 16271327ED78: Health status: FAILED!, Reported_Uncorrect is non-zero (245410), Reallocated_Event_Count is non-zero (1), Attribute...

`ata_errors` are added by this PR. `-s` self-test doesn't detect anything: ``` sudo /usr/sbin/smartctl -d ata -q silent -A /dev/sdb (debug) exit code: 0 (debug) zero exit code, status OK...

No problem. Currently the script executes following commands: 1. `smartctl -d ata -Hi /dev/sdb` 2. `smartctl -d ata -q silent -A /dev/sdb` 3. `smartctl -d ata -A /dev/sdb` 4. `smartctl...

Ok, then we should add more command executions.

@Napsty Now it's using `smarctl -a` and then checking the whole output. Is the attribute name `ata_errors` ok?

> On the other hand, it can also be very easy to miss it when we do become more fact dependent. When you add new feature, you should write tests...

having similar issue on TimescaleDB `2.24.0` with timestamp comparison: ``` WHERE "timestamp" > (now() - INTERVAL '14 days') ``` is there some known workaround?

Test failures seems to be unrelated. Upstream fixtures modules doesn't seem to support Puppet 7 anymore.