From mboxrd@z Thu Jan 1 00:00:00 1970 From: Artur Makowka Subject: Re: Trouble with 2.6.13.4-reiser4-4 Date: Wed, 02 Nov 2005 15:58:45 +0100 Message-ID: <4368D425.6060109@ursynow.2a.pl> References: <20051101015551.GA1059@148.Red-217-126-33.pooles.rima-tde.net.> <43670F5E.9030108@namesys.com> Reply-To: juice@ursynow.2a.pl Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <43670F5E.9030108@namesys.com> List-Id: Content-Type: text/plain; charset="windows-1252"; format="flowed" To: reiserfs-list@namesys.com Hans Reiser napisa=C5=82(a): >rvalles wrote: > > =20 > >>I've been using reiser4 for a while (since last december). Recently, >>I've upgraded from 2.6.12 to 2.6.13.4 patched with the reiser4 -4 patch >>available on the namesys ftp. >>For some reason, since using this kernel, every now and then the >>filesystem hits the disk really hard for 15 seconds or so blocking IO >>for a program or two (I guess those that triggered the problem) and >>slowing down IO seriously for others. It may be triggered by something >>as easy as pasting some text from firefox to a xterm running vim, and >>the text won't get pasted until the HD hitting ends. >>Just in case, I've booted a recent knoppix and ran the most recent fsck >>(thank you unionfs and apt-get) on the partition. It fixed some little >>stuff extents-related that didn't look serious at all. It didn't fix >>this, tho. >>Maybe it's worth noting that there are no reiser4 related errors or >>warnings, nor IO related, on syslog. >>I can provide more debug-related information on request. I'm looking >>forward for a patch against plain 2.6.14 (not -mm, that breaks stuff I >>need and is by itself far too unstable for my taste) to see if it fixes >>anything, too. >>Cheers. >>=20 >> >> =20 >> >We have a patch coming out any day now that will be a lot more stable >than what is currently the latest. > >Vitaly, please label our latest as not very stable in the meantime. > >hans > >. > > =20 > I have to agree with rvalles, as i also noticed very high LA and im=20 using latest reiser4 patch with 2.6.13.4 kernel. this is becouse of high=20 the I/O usage. I didnt know its reiser4 fault, but it seems logical (=20 other process are waiting for I/O and thats why so high LA ) becouse my machine had already high load anyways, so together with this=20 bug it caused few crashes, so im just wondering - how soon there is=20 going to be new patch, and is this issue already solved or beign worked on ? i know the patch will be ready when its ready ;) but i have free hosting=20 server and until its resolved i cant turn on all services. Also, there is nothing strange in logs. Also i have to say one good=20 thing about reiser4 - i had many crashes, some becouse of this bug, some=20 becouse server has very high load anyways, and reiser didnt delete or=20 corrupted even one file durning that. And i had to run --build-fs only=20 once or twice.