I reverted the recent changes to ObjectCacher. The problem is in flush_set: set<BufferHead*>::iterator next, it; next = it = dirty_or_tx_bh.begin(); while (it != dirty_or_tx_bh.end()) { next++; BufferHead *bh = *it; waitfor_commit.insert(bh->ob); bh_write(bh); it = next; } This resubmits IOs for writes that are already in flight. The fs client in particular calls this a lot, overwhelming the OSDs with dup IOs and stalling out completely. See /a/sage-2014-07-25_17:14:20-fs-wip-msgr-testing-basic-plana for a bunch of sample failures. Probably easy to fix, but simply reverting for now! sage -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html