I have an HP Proliant DL380 G4 that I have been ruing Ubuntu Server 12.04 on since 2012-12-22. I don't advertise it because I know that it is not that powerful and I run it at my house and don't want my ISP seeing a sudden influx of traffic. I have had no issues with it up until about a week ago. It has started having massive performance problems, so much so that text typed when logged in over ssh can take dozens of seconds to be shown. The configuration is as follows:
Hardware:
HP Proliant DL380 G4 (Quickspecs)
2x Single Core Hyper-Threaded Intel Xeon 3.6GHz (ark)
12GB RAM
First RAID with OS 3x 15k RPM 72.6GB SCSI OEM drives (RAID 0)
Second RAID 3x 10k RPM 72.6GB SCSI OEM drives (RAID 0)
OS:
Ubuntu 12.04.5 (Installed in 2012-12-22)
Primary Services:
LAMP
vsftpd
minecraft server(8GB max jvm allocation with openjdk 7)
The performance drops seems to come in waves about every minute. When it is ruing terribly there is no process listed in top using more cpu than java(minecraft server) at anytime. Java's typical cpu usage is between 37%-125% depending on the number of players(never more than 9). The problem causes the minecraft server to hang and crash after a few hours with a handful of players. The problem still exists when the minecraft server is not ruing so I have ruled it out as the cause. I do run websites on the server. They are intentionally not well known but are public facing.
I have run a memtest86+ scan and it passed. As I type this I am ruing a badblocks check on the 15k RPM RAID to check for HDD errors. I caot access individual drives or S.M.A.R.T. data behind the RAID controller that HP uses. I don't know if it is valid to run badblocks on a RAID 0 but I'm out of ideas.
There has been no gradual performance degradation that I have noticed. It was sudden and I caot determine why. I am at a loss as to where I should be looking to find the source of the problem. Any input would be greatly appreciated.
برچسب:
نویسنده: استخدام کار