Date: Sun, 11 Oct 1998 12:27:31 +0200 From: Thomas Zenker <thz@tuebingen.netsurf.de> To: current@FreeBSD.ORG Subject: system hangs on concurrent rm Message-ID: <19981011122731.40870@tuebingen.netsurf.de>
next in thread | raw e-mail | index | archive | help
After a running two concurrent "rm -rf " in obj tree the two processes locked up. This is on 3.0 current cvsupped at 7. Oct. With SOFTUPDATES enabled, including the changes of 3. of Oct. The disk this happened on is a SCSI disk (IBM DCAS). No other active processes were running, also disk activity only on this disk. I could switch vtys, but was unable to login. It seemed to me that both processes were blocked in the kernel, any process accessing filesystem or disk blocked immediately also. I left the system alone for a while, in case it there would have been a lockup on the SCSI bus it would have run into a timeout. Getting status with ^T showed me load avr going down to 0, after trying the tree finger magic to reboot, ^T showed a little load which went down quickly, but no disk activity. So did it the hard way, reset. It costed 4 hours of fsck to recover :-( but recovered :-) The problem with fsck was, that the lost+found directory got full quickly (nearly a whole obj tree), so fsck choked, only chance was fsck -y which was very slow obviously. Thomas Zenker To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19981011122731.40870>