48 ms·
I've seen many SSD failures in servers and all but one were just gone. No longer visible via the disk interface. The other one just had very slow response for q
by toast0 3y ago
I've seen many SSD failures in servers and all but one were just gone. No longer visible via the disk interface. The other one just had very slow response for quite some time (I think it was reallocating related to a large cluster of failing flash).
I have heard rumors that SSDs can go into read-only mode, but never experienced it. Our write rate was relatively low though, and we did not include it in our monitoring.
- coretx 3y agoIn such a scenario, iostat is your friend. https://manpages.debian.org/unstable/sysstat/iostat.1.en.html https://manpages.debian.org/unstable/sysstat/iostat.1.en.htm...