Date: Sat, 13 Feb 1999 10:44:35 -0500 From: "Clarence Brown" <clabrown@granitepost.com> To: <freebsd-fs@FreeBSD.ORG> Subject: uid 0 on /: file system full Message-ID: <000201be5767$c27ef220$8c6896d1@thunder.granitepost.com>
next in thread | raw e-mail | index | archive | help
I would appreciate any help, and apologize in advance if this is not the best place to post this, but it does involve the /procfs: On 2/11 I started getting MANY error messages of the following form: > pid 2499 (mail.local), uid 0 on /: file system full They are sent to the console and the logs, and occur ever couple minutes to 1/2 hour. The machine is running FreeBSD 2.1.1 and has been operating as an email server using sendmail and popper without problems for almost 2 years. Nothing significant (that I'm aware of) has changed. No changes to the init files, no new processes or daemons being run. I used the df command as follows to determine that /proc was what ran out of space. bash-2.00$ df Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 15957 42525 27% / /dev/wd0s1e 1873663 765002 958768 44% /usr procfs 4 4 0 100% /proc I searched the archives, and the only advice I found indicated that when procfs runs out of space you need to RECOMPILE your KERNEL with a larger number for MAXUSERS. This is very scary, but I wonder, SOMETHING must have changed. There are not more users on the system, and as far as I know, no new processes are running. If I can find out what changed, and just change it back, it seems like everything should return to normal error free operation. 1. What can cause procfs to run out of space? 2. Is procfs somehow affected by disk space? The only thing unusual I can see is my daily run output shows suddenly climbing disk usage (I don't know if this is a cause of the procfs error, or the effect of the log messages). I'll show them below: 8-30-98 (many months ago, notice almost same as 2-9) Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 16148 42334 28% / /dev/wd0s1e 1873663 492514 1231256 29% /usr 2-9-99 Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 15955 42527 27% / /dev/wd0s1e 1873663 521925 1201845 30% /usr 2-10-99 Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 15955 42527 27% / /dev/wd0s1e 1873663 521285 1202485 30% /usr 2-11-99 (sudden jump of 14% in /usr) Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 15955 42527 27% / /dev/wd0s1e 1873663 765215 958555 44% /usr 2-12-99 Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 15955 42527 27% / /dev/wd0s1e 1873663 762983 960787 44% /usr 2-13-99 (now sudden jump of 43% in /) Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/wd0a 63567 40965 17517 70% / /dev/wd0s1e 1873663 764968 958802 44% /usr I think all the logs are under /usr/var/log so if anything, it would seem that "/usr" should be growing steadily from all the error messages, but it only made the one time jump 30% to 44%. I think my user's home directories are under /home, so that would mean they are under the "/" above right? If there is "suddenly" a lot more mail that someone is leaving on the server, it might be taking up space in "/", but 43% seems like a lot. What would affect /procfs????? Thanks Cla. ps. I need to fix this by Monday. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000201be5767$c27ef220$8c6896d1>