Increased I/O utilization after new version deployment
Hello, maybe someone will have some ideas, because I'm running out of them.
New application version was deployed to production and we noticed increased C and D disk utilization. Reads/writes remains the same, but utilization skyrocketed. Couple days of not paying much attention we got total VM crash, CPU hit 100% and everything froze. Not sure about previous % Time in GC statistics, but current one looks awufly high.
Turning service off does not reduce utilization, tried waiting 30min. Only server reboot helps to fix it.
Application version changes:
1. Move from framework 4.5.2 to 4.8
2. Added support for PostgreSQL (previously only MSSQL was supported)
3. Small refactoring
4. Bug fixes
After tripple checking code, it does not look like refactoring or bug fixes could have caused something like that.
Grafana IO graph attached.
Performance monitor % Time in GC attached
Active time % keeps jumping 0 - 100%
5 Replies
What process is accessing the disk? What files are being accessed?
@canton7 As by windows disk activity. There are multiple instances deployed of same service. In theory all it should do is write log files.
Sort by the read / write columns, see what process is using the disk
It's very unstable, but ones on top are mostly
https://superuser.com/a/1753741 ? Maybe related to adding Postgres to your stack?
Super User
System keeps writing in $LogFile and $Mft on external drives, cau...
My external drives kept randomly spinning up for no reason. I disabled indexing and optimizing on them and did all other commonly recommended things, but it kept happening.
Then I installed Process