Performance tuning is the process of modifying a system to improve its efficiency. This is also known as optimizing.
16
votes
3answers
4k views
Linux - real-world hardware RAID controller tuning (scsi and cciss)
Most of the Linux systems I manage feature hardware RAID controllers (mostly HP Smart Array). They're all running RHEL or CentOS.
I'm looking for real-world tunables to help optimize performance for ...
2
votes
1answer
1k views
Openfiler iSCSI performance
Hoping someone can point me in the right direction with some iSCSI performance issues I'm having. I'm running Openfiler 2.99 on an older ProLiant DL360 G5. Dual Xeon processor, 6GB ECC RAM, Intel ...
80
votes
4answers
90k views
FreeBSD performance tuning. Sysctls, loader.conf, kernel
I wanted to share knowledge of tuning FreeBSD via sysctl.conf / loader.conf
/ KENCONF / etc. It was initially based on Igor Sysoev's (author of nginx)
presentation about FreeBSD tuning up to ...
13
votes
5answers
19k views
yslow says etags are misconfigured. how to configure etags properly on IIS7?
running ySlow on a website I support, I noticed it reported that the etags are misconfigured for the site images (e.g. *.jpg, *.png, *.gif). Can anyone help explain what I have to do to get IIS7 ...
3
votes
2answers
513 views
CPU0 is swamped with eth1 interrupts
I've got an Ubuntu VM, running inside Ubuntu-based Xen XCP. It hosts a custom FCGI-based HTTP service, behind nginx.
Under load from ab the first CPU core is saturated, and the rest is under-loaded.
...
3
votes
3answers
954 views
Linux's best filesystem to work with 10000's of files without overloading the system I/O
It is known that certain AMD64 Linuxes are subject of being unresponsive under heavy disk I/O (see Gentoo forums: AMD64 system slow/unresponsive during disk access (Part 2)), unfortunately have such ...
7
votes
13answers
5k views
Postfix performance
Running postfix on ubuntu, sending alot of mail ( ~ 1 million messages ) per day. loads are extremly high but not much in terms of cpu and memory load. Anyone in a similiar situation and know how to ...
18
votes
7answers
6k views
Is it OK to run perfmon on production servers? And why?
Or should perfmon be limited to a Dev/QA server with load tests that simulate production activity?
I'd like to run perfmon for two days (like Sql Server master Brent Ozar suggests) to get an overall ...
7
votes
1answer
501 views
Why change net.inet.tcp.tcbhashsize in FreeBSD?
In virtually every FreeBSD network tuning document I can find:
# /boot/loader.conf
net.inet.tcp.tcbhashsize=4096
This is usually paired with some unhelpful statement like "TCP control-block hash ...
5
votes
8answers
6k views
Hard Drive Speed: RPM vs. Cache
When trying to choose a SATA magnetic hard drive (not SSD) for high performance for both random and linear access, which should be the primary factor?
For example: would a 10k RPM drive with 16MB ...
4
votes
1answer
423 views
Ubuntu HTTP latency falling into strange quantiles
I have an Ubuntu 10.10 server with plenty of RAM, bandwidth and CPU. I'm seeing a strange, repeatable pattern in the distribution of latencies when serving static files from both Apache and nginx. ...
4
votes
7answers
1k views
Linux tool/command to check file system performance
What is the simplest way to check read/write performance to a specific location (e.g. a mounted iSCSI device).
I suspect I can't use hdparm because that's lower level. Am I right?
2
votes
3answers
4k views
How can I improve NFS performance on a gigabit LAN?
I have a Solaris 10 server with a Mac OS X clients on a gigabit LAN. The Solaris server exports a RAID-Z backed NFS share and all the clients connect to it. However, transferring files is slower ...
4
votes
1answer
521 views
How do I test the effect of ionice (against a device using the cfq scheduler)?
I'm trying to construct an experiment to measure the effect of ionice. What I'd like to do is to (per another answer on serverfault) cause frequent enough I/O that a sufficiently "niced" process is ...
3
votes
1answer
544 views
Disabling CPU management
If I add the following processor.max_cstate=0 to the kernel command line for boot up, does that disable all CPU power management and throttling?
I also found:
...
3
votes
2answers
194 views
Uneven CPU core utilization
tl;dr: First CPU core is consistently saturated, all other cores are consistently under-loaded.
A VM, inside Ubuntu-based Xen XCP:
$ uname -a
Linux MYHOST 2.6.38-15-virtual #59-Ubuntu SMP Fri Apr ...
3
votes
3answers
1k views
HP P410 RAID + Samsung 830 SSDs + Debian 6.0 - What performance to expect?
I'm renting two dedicated servers from a hosting company. Here are the specs:
server1:
HP ProLiant DL165 G7
2x AMD Opteron 6164 HE 12-Core
40 GB RAM
HP Smart Array P410 RAID controller
2x Samsung 830 ...
2
votes
1answer
196 views
How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
The ionice man page says
A program running with idle io priority will only get disk time
when no other program has asked for disk io for a defined grace period.
Where is this "grace period" ...
1
vote
1answer
72 views
How do I disable “Processor Power and Utilization Monitoring” on an HP ProLiant server?
I'm tuning my server for low-latency The guide suggests to disable this option "Processor Power and Utilization Monitoring" because:
Disables iLO Processor State Mode Switching
and Insight Power ...
1
vote
1answer
544 views
Lot of FIN_WAIT2, CLOSE_WAIT , LAST_ACK and TIME_WAIT in Haproxy
We are running haproxy in production for around 10k+ concurrent users . But we are seeing lot of FIN_WAIT2, CLOSE_WAIT , LAST_ACK and TIME_WAIT in the netstat output. This output is on a 8G ...
1
vote
2answers
395 views
Troubleshooting poor disk IO performance
Note that while this question is a little Redis-specific, the main problem is generic: a process takes so much HDD IO write bandwidth that other processes can't write anything.
We've got an Ubuntu VM ...
0
votes
5answers
1k views
Very low performance of raid6
I noticed that writing to raid-6 is very low, but when I make tests with hdparm the speed is reasonable:
dd if=/dev/zero of=/store/01/test.tmp bs=1M count=10000
Gives: 50Mb/s or even less.
The ...
