public inbox for cluster-cvs@sourceware.org help / color / mirror / Atom feed
From: jbrassow@sourceware.org To: cluster-cvs@sources.redhat.com Subject: cluster/cmirror-kernel/src dm-cmirror-client.c ... Date: Wed, 14 Mar 2007 04:28:00 -0000 [thread overview] Message-ID: <20070314042832.27983.qmail@sourceware.org> (raw) CVSROOT: /cvs/cluster Module name: cluster Branch: RHEL4 Changes by: jbrassow@sourceware.org 2007-03-14 04:28:32 Modified files: cmirror-kernel/src: dm-cmirror-client.c dm-cmirror-server.c Log message: Bug 231230: leg failure on cmirrors causes devices to be stuck in SUSPE... The problem here appears to be timeouts related to clvmd. During failures under heavy load, clvmd commands (suspend/resume/ activate/deactivate) can take a long time. Clvmd assumes to quickly that they have failed. This results in the fault handling being left half done. Further calls to vgreduce (by hand or by dmeventd) will not help because the _on-disk_ version of the meta-data is consistent - that is, the faulty device has been removed. The most significant change in this patch is the removal of the 'is_remote_recovering' function. This function was designed to check if a remote node was recovering a region so that writes to the region could be delayed. However, even with this function, it was possible for a remote node to begin recovery on a region _after_ the function was called, but before the write (mark request) took place. Because of this, checking is done during the mark request stage - rendering the call to 'is_remote_recovering' meaningless. Given the useless nature of this function, it has been pulled. The benefits of its removal are increased performance and much faster (more than an order of magnitude) response during the mirror suspend process. The faster suspend process leads to less clvmd timeouts and reduced probability that bug 231230 will be triggered. However, when a mirror device is reconfigured, the mirror sub-devices are removed. This is done by activating them cluster-wide before their removal. With high enough load during recovery, these operations can still take a long time - even though they are linear devices. This too has the potential for causing clvmd to timeout and trigger bug 231230. There is no cluster logging fix for this issue. The delay on the linear devices must be determined. A temporary work-around would be to increase the timeout of clvmd (e.g. clvmd -t #). Patches: http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/cmirror-kernel/src/dm-cmirror-client.c.diff?cvsroot=cluster&only_with_tag=RHEL4&r1=1.1.2.40&r2=1.1.2.41 http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/cmirror-kernel/src/dm-cmirror-server.c.diff?cvsroot=cluster&only_with_tag=RHEL4&r1=1.1.2.25&r2=1.1.2.26
next reply other threads:[~2007-03-14 4:28 UTC|newest] Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top 2007-03-14 4:28 jbrassow [this message] -- strict thread matches above, loose matches on Subject: below -- 2007-10-03 19:02 jbrassow 2007-09-27 20:31 jbrassow 2007-09-26 3:15 jbrassow 2007-09-21 20:07 jbrassow 2007-09-13 15:24 jbrassow 2007-07-11 16:18 jbrassow 2007-04-26 16:55 jbrassow 2007-04-26 16:54 jbrassow 2007-04-24 20:10 jbrassow 2007-04-24 20:08 jbrassow 2007-04-10 7:13 jbrassow 2007-04-10 7:12 jbrassow 2007-04-05 21:33 jbrassow 2007-04-05 21:32 jbrassow 2007-04-03 18:23 jbrassow 2007-04-03 18:21 jbrassow 2007-03-22 22:34 jbrassow 2007-03-22 22:22 jbrassow 2007-02-26 17:38 jbrassow 2007-02-20 19:35 jbrassow 2007-02-19 16:29 jbrassow 2007-02-14 17:44 jbrassow 2007-02-02 17:22 jbrassow 2007-01-08 19:28 jbrassow 2006-12-07 18:58 jbrassow 2006-09-05 17:50 jbrassow 2006-09-05 17:48 jbrassow 2006-07-27 23:11 jbrassow 2006-07-27 23:11 jbrassow 2006-07-22 22:19 jbrassow 2006-07-22 22:19 jbrassow 2006-07-22 22:12 jbrassow 2006-06-29 19:49 jbrassow 2006-06-29 19:48 jbrassow 2006-06-29 19:46 jbrassow 2006-06-27 20:19 jbrassow 2006-06-15 19:48 jbrassow 2006-06-15 19:34 jbrassow 2006-06-13 16:26 jbrassow 2006-06-02 22:04 jbrassow 2006-05-15 16:26 jbrassow 2006-02-08 21:59 jbrassow 2006-02-08 16:03 jbrassow 2006-02-06 23:04 jbrassow 2006-02-06 22:58 jbrassow 2006-02-06 18:14 jbrassow 2005-10-27 21:27 jbrassow 2005-10-27 11:03 jbrassow 2005-08-11 18:26 jbrassow 2005-07-28 22:53 jbrassow
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20070314042832.27983.qmail@sourceware.org \ --to=jbrassow@sourceware.org \ --cc=cluster-cvs@sources.redhat.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).