Messages in this thread | | | From | "Rob Mueller" <> | Subject | Re: Strange load spikes on 2.4.19 kernel | Date | Sat, 12 Oct 2002 16:52:31 +1000 |
| |
> It commits your changes to the journal every five seconds. But your data > is then only in the journal. It still needs to be written into your files. > That writeback is controlled by the normal kernel 30-second writeback > timing. If that writeback isn't keeping up, kjournald needs to > force the writeback so it can recycle that data's space in the journal. > > While that writeback is happening, everything tends to wait on it.
Doesn't bdflush let you control this? I noted in my first post that we'd played with changing bdflush params as described here:
http://www-106.ibm.com/developerworks/linux/library/l-fs8.html?dwzone=linux
And set them to this:
[root@server5 hm]# cat /proc/sys/vm/bdflush 39 500 0 0 60 300 60 0 0
Shouldn't this reduce it to writing every 3 seconds? We tried lowering some of the values even further based on the description here:
http://www.bb-zone.com/zope/bbzone/docs/slgfg/part2/cha04/sec04
So we altered the first one (nfract) to 10(%) to try and keep the dirty buffer list small, but that didn't help either. I sort of thought that age_buffer at 3 seconds would be more likely to activate anyway than 40% of buffers being dirty?
> It is suspected that ext3 gets the flushtime on those buffers > wrong as well, so the writeback isn't happening right.
So you're saying that ext3 is somehow breaking the standard kernel writeback code? Is this something they know about, and/or are addressing?
Rob
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |