Date: Thu, 13 Nov 2014 15:20:50 -0500 From: Jim Mercer <jim@reptiles.org> To: freebsd-fs@freebsd.org Subject: Re: leaking lots of unreferenced inodes (pg_xlog files?), maybe after moving tables and indexes to tablespace on different volume Message-ID: <20141113202050.GA45077@reptiles.org> In-Reply-To: <201303160401.r2G41Um7026132@chez.mckusick.com> References: <51439243.5020604@FreeBSD.org> <201303160401.r2G41Um7026132@chez.mckusick.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Mar 15, 2013 at 09:01:30PM -0700, Kirk McKusick wrote: > I don't know how, but somehow something is holding references to > the removed files causing them to fail to be reclaimed. > > Could you run your system for a while to build up a new set of > these files, then run a script with the `df -ih' as before. Then > run `vmstat -m', `sysctl debug', and fstat -f /usr' both before > and after doing the umount/mount. Hopefully that will give us > some more clues as to what is happening. i am seeing a similar behaviour: FreeBSD db02 9.2-PRERELEASE FreeBSD 9.2-PRERELEASE #0: Mon Aug 12 18:40:42 PKT 2013 root@rms-db02.telenor.innexiv.com:/usr/obj/usr/src/sys/INXVMPG amd64 postgresql-server-9.2.4 however, i'm unsure its xlog files in my case. i have postgres installed on /home root@rms-db02 /var/log]# df -h Filesystem Size Used Avail Capacity Mounted on /dev/da0p3 11G 3.2G 7.5G 30% / devfs 1.0k 1.0k 0B 100% /dev /dev/da1p1 1.5T 671G 754G 47% /home procfs 4.0k 4.0k 0B 100% /proc [root@rms-db02 /var/log]# ls -ld /usr/local/pgsql lrwxr-xr-x 1 root wheel 11 Aug 12 2013 /usr/local/pgsql -> /home/pgsql in my case, it over the course of 17 days (it seems to be happening faster now, at one point it was 6 months or more, now sooner). the "Used" space on root goes up to approx 95% at which point i halt the applications and reboot. i hunted around with lsof and such, and had another guy also poke about, and neither of us could find the place where the space was being used. i wanted to follow through on the diagnostic you provided, but since it was the root filesystem, it wasn't straightforward. best i could do was a 'df -i', which showed, immediately after reboot, into single user, before fsck, that the space had been recovered. comparing the inode count between singleuser and booted up multiuser, the booted up multiuser had +9 inodes, so, no vast difference. i am going to be upgrading to more recent freebsd and postgres soon, but, in the meantime, was there any further investigation of this? --jim -- Jim Mercer Reptilian Research jim@reptiles.org +1 416 410-5633 "He who dies with the most toys is nonetheless dead"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20141113202050.GA45077>