4 ms·
Anyone has an idea if there is any way to distinguish random from sequential IO? In my experience sequential IO is never the problem. Instead, it seems to me t
by sprachspiel 17y ago
Anyone has an idea if there is any way to distinguish random from sequential IO?
In my experience sequential IO is never the problem. Instead, it seems to me that random seeks are really the only performance problem nowadays. In the most extreme case throughput is only ~100Bits/s instead of ~100MB/s. Unfortunately random seeks are hidden behind abstraction layers and thus quite invisible to programmers (until the system freezes). Maybe we just need to wait for SSDs to become cheap.
- voberoi 17y agoTake a look at blktrace on Linux. It logs all block I/O layer activity from when an I/O request is sent to the disk to when it's serviced. It also outputs which sectors on the disk were written to service the request. People have written a few tools to read these logs and show meaningful numbers/pretty graphs -- btt is the official one, and here's another one that I like: http://oss.oracle.com/~mason/seekwatcher/ http://oss.oracle.com/~mason/seekwatcher/. Among other things, you can distinguish random from sequential I/O using blktrace and its accompanying set of tools by looking at the distance between subsequent I/O requests sent to the disk.
- gnaritas 17y agoSSD's are cheap.
- jacquesm 17y agoWould you care to quantify that ?
- gnaritas 17y agoI picked up my Intel (the good controller, MLC version) for $400, best upgrade ever, that's fucking cheap.
- blasdel 17y ago...and cheap SSDs are pathologically bad at random writes -- at least an order of magnitude worse than an mediocre hard drive on average, with regular latency spikes of several seconds!
- wmf 17y agoWhen measured in $/IOPS, even good SSDs are cheap. I think the real problem is the lack of caching software (but I would think that).