Tag Info

New answers tagged

0

This is an old question, but we faced the same issues (High IO waits, and terrible insert / update speeds) the past week on a new dedicated server and this solution addresses this issue directly. Disabling journaling with tune2fs -O "^has_journal" /dev/<drive> was the quickest solution as it eliminates the IO wait because of the JDB2 process. But this ...


1

Yes, there is. Using a tap device causes context switches between the client process and the process that holds the tap device. It also causes additional copies, as data needs to be copied from the client process into the kernel, and then from the kernel into the tap holder. However, Linux is pretty good at context switching, and in-cache copies are ...


1

I'm posting this as an answer due to insufficient commenting permissions. From my experience VPN over tcp - tap was abysmal at best over wireless connections due to the the nature of TCP. When switching to UDP no more connection timeouts occurred - even better with tun device. I would suggest you test under poor network connections both TCP and UDP and ...


1

Look into using background jobs for the Get-ScheduledTask cmdlet. This will run the command asynchronously. foreach ($computer in $computers) { start-job -scriptblock { #Parse Args $CompName = $args[0] $TaskName = $args[1] Get-ScheduledTask -TaskName $TaskName -CimSession (New-CimSession -ComputerName $CompName) } ...


3

Brandan Gregg has made a number of visual diagrams of which tool to use on what part of your system that make really nice cheat sheets: Tools: Which options to use in sar to monitor: And closely related: Translate observations into tuning with:


0

I want to add some tool to helps for get accurate result 1.Lsof 2.Psacct or Acct – Monitor User Activity 3.Monit – Linux Process and Services Monitoring 4.NetHogs 5.Monitorix – System and Network Monitoring 6.Arpwatch – Ethernet Activity Monitor 7.Suricata – Network Security Monitoring 8.VnStat PHP – Monitoring Network Bandwidth 9.Nagios – ...


2

There is no clear cut answer to such question, however I will provide you with some links that should set you on the right path of optimizing the VM setup. I would start by reading the official documentation from Microsoft regarding performance: Performance Guidance for SQL Server in Azure Virtual Machines: ...


0

tunefs have an option -S -S size Specify the softdep journal size in bytes. The minimum is 4M. Seems that default value is to small for your FS. You have to fsck your partition, then disable journaling with tunefs -j disable /dev/da0p0x and then reenable journaling with bigger journal tunefs -j enable -S 134217728 /dev/da0p0x


2

The first thing you need to look at is to determine where your bottleneck is. Is the system's CPU pegged, or is there CPU available but the web application's processes aren't using it effectively (ie, not enough worker threads)? Is the system out of RAM, or doing a lot of swapping? Are you bottlenecking on your database? Do you have slow queries, need ...


2

If you use the latest stable version of proftpd (1.3.5) then there is bug when using TLS: http://bugs.proftpd.org/show_bug.cgi?id=4108 TJ Saunders 2014-09-24 02:46:33 UTC Pull request opened which fixes the issue: https://github.com/proftpd/proftpd/pull/48 Short version: we do NOT want the TCP_CORK/TCP_NOPUSH socket option enabled on ...


1

Look into the buffer command, basically it speeds allows simultaneous read/write activity instead of the normal read, then write, then read, etc.... Basically it does this by spawning two sub-processes. The processes communicate using a shared memory buffer. The command line parameters are similar to dd with the addition of parameters to size the shared ...


0

The dedicated server has 250Mbps DL speed and 150Mbps UP speed My home internet has 100Mbps DL speed and 7Mbps UP speed So the link between the two is theoretically maxing at 10MiB/s. Well, your circuit at work and your circuit at home are only a small piece of the whole system. When you purchase a circuit that will provide X Mbps download or X Mbps ...


1

JBOD used generally means exactly what it says on the tin - just a bunch of disks. There's no raid, no disk consolidation, nothing. You'll see a whole bunch of separate devices down your controller. Intel's 'JBOD' unit needs a RAID adaptor if you want to do anything more clever. For example: ...


2

A JBOD (Just a Bunch Of Disks) is just that. It has no controller aside from what is necessary to provide power and data lanes to the disks. In a case of a more sophisticated JBOD, you will see a backplane and, depending on the technology used, maybe an expander. The controller needs to be placed on the host's side, it will be connected to the JBOD with ...


-1

TCP/IP stack on linux is already very optimized and typically nothing is needed. For instance, setting local port range to get a few extra ports is almost certainly not needed. In terms of time/wait being bad, it is just part of using tcp. If you really want to have less ports in that state, change tcp_fin_timeout or tcp_keepalive values. Although you ...


25

CPU time is allocated in discrete time slices (ticks). For a certain number of time slices, the cpu is busy, other times it is not (which is represented by the idle process). In the picture below the CPU is busy for 6 of the 10 CPU slices. 6/10 = .60 = 60% of busy time (and there would therefore be 40% idle time). A percentage is defined as "a number or ...


10

The CPU time is the time that the process is using the CPU - converting it to a percentage is done by dividing by the amount of real time that's passed. So, if I have a process that uses 1 second of CPU time over a period of 2 seconds, it's using 50% of a CPU. In the case of your MATLAB process, 217% indicates that it's used 2.17 seconds of CPU time per ...


2

DNS operates best in a Master/Slave scenario. Consider this: you have two nameservers, ns1 and ns2. To help balance the load, you have two NS records in your domain record for both ns1 and ns2. By the very nature of DNS, clients will query the nameservers in random order because nameserver lookups are returned in a random order. See this answer for helpful ...


0

There could be many reasons for this ranging from network configured incorrectly to the source being the issue. However assuming that the data is located locally or on a fast server I would look at the following items. I have just built a new Dell server this week and this really helped (assuming HP has the same options) 1) DO NOT enable RTID - ...


0

I have solved this on my own after a bunch more research. I had a rewrite script set that I forgot I set, and said script was crashing on 90% of URLs. Fixing the script fixed this issue.


1

CPL is load average figures reflecting the number of threads that are available to run on a CPU (i.e. part of the runqueue) or that are waiting for disk I/O. You seem to have ~16 processes that seems to be waiting for the disk. That's the reason you see the cpu mostly idle, it doesn't have anything to do other than waiting for the disk. I would check the ...


1

Keep in mind that harddrives have a certain sector size. Writing normally goes per sector, so you should expect lower speeds below the sector size. In additionn most raid configuration either apply mirroring (same data on multiple disks, speeeds up reading as information only needs to be read from 1 disk instead of X disks, however writes have to go to all ...


2

Based on the documentation you quoted I find it quite clear that one is covering the entire duration from one system call to the next, while the other covers only the time within a system call. The percentage of time spent inside system calls versus the percentage of time spent outside system calls will roughly tell you if a process is CPU bound. A CPU ...


2

The time outside the system call is the time spent running your program's code before it gets to the next system call.



Top 50 recent answers are included