Date: Tue, 24 Apr 2001 11:05:31 -0400 (EDT) From: Marius <marius@mail.communityconnect.com> To: questions@FreeBSD.org Cc: marius@mail.communityconnect.com Subject: nfs stuck in objtrm Message-ID: <Pine.BSF.4.21.0104241043140.23218-100000@utterlux.hq.communitconnect.com>
next in thread | raw e-mail | index | archive | help
My company has a number of FreeBSD servers, 3.5-STABLE, that run thttpd to serve static content. The staic content is stored on nfs mounts. In general our FreeBSD servers are very fast, very efficient, they are the real workhorses of our server farm. We have two main problems that keep re-ocurring. One is that machines occationaly get stuck in vmopar. (Kernel Bug 14033) A problems which I believe is still unresolved. The second, more frequent annoyance is thttpd getting stuck in the 'objtrm' state while the nfs client hangs is vm_object_terminate(). Searching the archives I see this was once Kern/1159 but obviously that was a while ago. All my searches so far seem to indicate that is problem should have been cleared up long ago, yet I have had to reboot 3 servers in the last 12 hours because of it. The only thing we can do when this happens is ungracefully reboot the system without syncing the file system, as a normal reboot will just hang the machine as the objtrm process just won't die. I wish I knew a better way or could help to create a better resolution. Has anyone else seen this behavior recently? I will hopefully soon upgrade at least a test server to to the latest 4.x stable as well as the most recent release of thttpd so I can take full advantage of accept filters. Hopefully that will solve/help some of our woes. But I was wondering if anyone had any insite/experiece similar to ours with objtrm. -Marius M. Rex To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0104241043140.23218-100000>