| Message ID | or62qjxzq3.fsf@livre.localdomain (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, 12 Apr 2011, Alexandre Oliva wrote: > I'm getting relatively frequent MDS crashes. I believe this happens as > a replay-standby node is about to take over once the MDS it follows is > restarted, most often because it stopped making progress for a long > time. In the core files I checked, waitfor_trim was empty, so this > shouldn't have ill effects. Since I'm not sure how to trigger this > particular assertion failure, I can't say that I've tested this patch > thoroughly, but unless the purpose of the assertion is to catch the == > case, I think it's safe to put it in. The assertion is that way on purpose.. we shouldn't be trimming to a offset we've already trimmed to. Looking at the code it's not obvious how we're getting into that particular corner, though. Can you reproduce the problem with debug journaler = 20 ? That should have enough information to track down exactly where things are going awry. Thanks! sage > > --- > Relax journal trim_finish assertion to pass when (to == trimmed_pos). > > Signed-off-by: Alexandre Oliva <oliva@lsd.ic.unicamp.br> > --- > src/osdc/Journaler.cc | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/src/osdc/Journaler.cc b/src/osdc/Journaler.cc > index f3241c4..4f4811c 100644 > --- a/src/osdc/Journaler.cc > +++ b/src/osdc/Journaler.cc > @@ -960,7 +960,7 @@ void Journaler::_trim_finish(int r, uint64_t to) > assert(r >= 0); > > assert(to <= trimming_pos); > - assert(to > trimmed_pos); > + assert(to >= trimmed_pos); > trimmed_pos = to; > > // finishers? > -- > 1.7.4.2 > > -- > Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/ > You must be the change you wish to see in the world. -- Gandhi > Be Free! -- http://FSFLA.org/ FSF Latin America board member > Free Software Evangelist Red Hat Brazil Compiler Engineer > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/src/osdc/Journaler.cc b/src/osdc/Journaler.cc index f3241c4..4f4811c 100644 --- a/src/osdc/Journaler.cc +++ b/src/osdc/Journaler.cc @@ -960,7 +960,7 @@ void Journaler::_trim_finish(int r, uint64_t to) assert(r >= 0); assert(to <= trimming_pos); - assert(to > trimmed_pos); + assert(to >= trimmed_pos); trimmed_pos = to; // finishers?
I'm getting relatively frequent MDS crashes. I believe this happens as a replay-standby node is about to take over once the MDS it follows is restarted, most often because it stopped making progress for a long time. In the core files I checked, waitfor_trim was empty, so this shouldn't have ill effects. Since I'm not sure how to trigger this particular assertion failure, I can't say that I've tested this patch thoroughly, but unless the purpose of the assertion is to catch the == case, I think it's safe to put it in. --- Relax journal trim_finish assertion to pass when (to == trimmed_pos). Signed-off-by: Alexandre Oliva <oliva@lsd.ic.unicamp.br> --- src/osdc/Journaler.cc | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)