Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 02/18/2014 11:24 AM, Alex Lyakas wrote: > Hi Josef, is this the commit to look at: > 6df9a95e63395f595d0d1eb5d561dd6c91c40270 Btrfs: make the chunk > allocator completely tree lockless > > or some other commits are also relevant? > It's been so long but I'm pretty sure everything you need is in that patch. Thanks, Josef -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJTA4nPAAoJEANb+wAKly3BlNwP/AyG7LtDo6YYaYvXJyBJa1Vj hq2C48lwQhSjAYfn5QJ05AUOgL4xAb1THSjDqTIyoyEXGwBnRLEaX3/MFygthrxi 9u137ys1C/EQr3fmRecdz6Qpojkf5EAxiK8J0nL+G/BXoJYdwCYUj4oLOgqwP6/X /XhsqyCLmj8jATndYCz7Z68xfutF37xtId0mWEsRrnvMqrT5nDvA/WpkzYE+ovc3 OhffFHfHJAf94qMb6EtSpH3E2MJDIYfp6cIAEgEK2ougZLnf0lkjcCXd2B6fRLcY 9WuZaVsi4J+vqGxVwnxDaJ7TbjEDXbl+bnAs5R5VDKZUy56zOxNA9//ejCuYtl/P r5K0PKZXu81wiK22DbF0hhZfzdkElnVqx8DSgwTyyo5aJTj6cNuDRdPmTz4TEbib N8z7rGC85Y4Z9Z1Gwnj3cD6pKQU4+anUhkIWNFVM9SpWbjYXgjjTMAj/LaM6GhJL OptTORUwu4+9hGnfu7ItL8uyVrBwyh9cUcbru79D0+YyyWR5fDsgYFCtvUuhJ16q vrViGT2MVyt4ZevvJMG02997sC8OCyeF4W0eQgyvgSOJToeoOJ57j8z/mSUntqDE 94f6hqOBjN6UY6/2FFILeMH0xuF0Li5JUOYB5Da99iHByeHQ4hrBWVyyvZfqW4vN YY32d8J7Ine1N7/IZdVh =jn5g -END PGP SIGNATURE- -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
Hi Josef, is this the commit to look at: 6df9a95e63395f595d0d1eb5d561dd6c91c40270 Btrfs: make the chunk allocator completely tree lockless or some other commits are also relevant? Alex. On Tue, Feb 18, 2014 at 6:06 PM, Josef Bacik wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > > > On 02/18/2014 10:47 AM, Alex Lyakas wrote: >> Hello Josef, >> >> On Tue, Dec 18, 2012 at 3:52 PM, Josef Bacik >> wrote: >>> On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: An user reported that he has hit an annoying deadlock while playing with ceph based on btrfs. Current updating device tree requires space from METADATA chunk, so we -may- need to do a recursive chunk allocation when adding/updating dev extent, that is where the deadlock comes from. If we use SYSTEM metadata to update device tree, we can avoid the recursive stuff. >>> >>> This is going to cause us to allocate much more system chunks >>> than we used to which could land us in trouble. Instead let's >>> just keep us from re-entering if we're already allocating a >>> chunk. We do the chunk allocation when we don't have enough >>> space for a cluster, but we'll likely have plenty of space to >>> make an allocation. Can you give this patch a try Jim and see if >>> it fixes your problem? Thanks, >>> >>> Josef >>> >>> >>> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c >>> index e152809..59df5e7 100644 --- a/fs/btrfs/extent-tree.c +++ >>> b/fs/btrfs/extent-tree.c @@ -3564,6 +3564,10 @@ static int >>> do_chunk_alloc(struct btrfs_trans_handle *trans, int >>> wait_for_alloc = 0; int ret = 0; >>> >>> + /* Don't re-enter if we're already allocating a chunk */ >>> + if (trans->allocating_chunk) + return >>> -ENOSPC; + space_info = __find_space_info(extent_root->fs_info, >>> flags); if (!space_info) { ret = >>> update_space_info(extent_root->fs_info, flags, @@ -3606,6 +3610,8 >>> @@ again: goto again; } >>> >>> + trans->allocating_chunk = true; + /* * If we have mixed >>> data/metadata chunks we want to make sure we keep * allocating >>> mixed chunks instead of individual chunks. @@ -3632,6 +3638,7 @@ >>> again: check_system_chunk(trans, extent_root, flags); >>> >>> ret = btrfs_alloc_chunk(trans, extent_root, flags); + >>> trans->allocating_chunk = false; if (ret < 0 && ret != -ENOSPC) >>> goto out; >>> >>> diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c >>> index e6509b9..47ad8be 100644 --- a/fs/btrfs/transaction.c +++ >>> b/fs/btrfs/transaction.c @@ -388,6 +388,7 @@ again: >>> h->qgroup_reserved = qgroup_reserved; h->delayed_ref_elem.seq = >>> 0; h->type = type; + h->allocating_chunk = false; >>> INIT_LIST_HEAD(&h->qgroup_ref_list); >>> INIT_LIST_HEAD(&h->new_bgs); >>> >>> diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h >>> index 0e8aa1e..69700f7 100644 --- a/fs/btrfs/transaction.h +++ >>> b/fs/btrfs/transaction.h @@ -68,6 +68,7 @@ struct >>> btrfs_trans_handle { struct btrfs_block_rsv *orig_rsv; short >>> aborted; short adding_csums; + bool allocating_chunk; enum >>> btrfs_trans_type type; /* * this root is only needed to validate >>> that the root passed to >> >> I hit this problem in a following scenario: - a data chunk >> allocation is triggered, and locks chunk_mutex - the same thread >> now also wants to allocate a metadata chunk, so it recursively >> calls do_chunk_alloc, but cannot lock the chunk_mutex => deadlock - >> btrfs has only one metadata chunk, the one that was initially >> allocated by mkfs, it has: total_bytes=8388608 bytes_used=8130560 >> bytes_pinned=77824 bytes_reserved=180224 so bytes_used + >> bytes_pinned + bytes_reserved == total_bytes >> >> Your patch would have returned ENOSPC and avoid the deadlock, but >> there would be a failure to allocate a tree block for metadata. So >> the transaction would have probably aborted. >> >> How such situation should be handled? >> >> Idea1: - lock chunk mutex, - if we are allocating a data chunk, >> check whether the metadata space is below some threshold. If yes, >> go and allocate a metadata chunk first and then only a data chunk. >> >> Idea2: - check if we are the same thread that already locked the >> chunk mutex. If yes, allow recursive call but don't attempt to >> lock/unlock the chunk_mutex this time >> >> Or some other way? >> > > I fixed this with the delayed chunk allocation stuff which doesn't > actually do the block group creation stuff until we end the > transaction, so we can allocate metadata chunks without any issue. > Thanks, > > Josef > -BEGIN PGP SIGNATURE- > Version: GnuPG v1 > Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ > > iQIcBAEBAgAGBQJTA4UMAAoJEANb+wAKly3B+KEP/RdlEyJWydetjQxllF0cgHY1 > UraqWBl+mSSHlwZlHyGjmAu6cK6n+QfTZtdIBhihdY50UcvMuWtVmz2JzlbxeO5+ > 88dBevADmW+QQoRl0yyQgnjlLWm+LvMTgOd1r+DZqlGs6sdX05dMI207+fQOW+c4 > P+UKbT/eUYRVC4K//J1GUk4Yh3Q70U25321RWCehSUciwDVJO2LztD9VBAgh3qUc > o5uh5sysh
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 02/18/2014 10:47 AM, Alex Lyakas wrote: > Hello Josef, > > On Tue, Dec 18, 2012 at 3:52 PM, Josef Bacik > wrote: >> On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: >>> An user reported that he has hit an annoying deadlock while >>> playing with ceph based on btrfs. >>> >>> Current updating device tree requires space from METADATA >>> chunk, so we -may- need to do a recursive chunk allocation when >>> adding/updating dev extent, that is where the deadlock comes >>> from. >>> >>> If we use SYSTEM metadata to update device tree, we can avoid >>> the recursive stuff. >>> >> >> This is going to cause us to allocate much more system chunks >> than we used to which could land us in trouble. Instead let's >> just keep us from re-entering if we're already allocating a >> chunk. We do the chunk allocation when we don't have enough >> space for a cluster, but we'll likely have plenty of space to >> make an allocation. Can you give this patch a try Jim and see if >> it fixes your problem? Thanks, >> >> Josef >> >> >> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c >> index e152809..59df5e7 100644 --- a/fs/btrfs/extent-tree.c +++ >> b/fs/btrfs/extent-tree.c @@ -3564,6 +3564,10 @@ static int >> do_chunk_alloc(struct btrfs_trans_handle *trans, int >> wait_for_alloc = 0; int ret = 0; >> >> + /* Don't re-enter if we're already allocating a chunk */ >> + if (trans->allocating_chunk) + return >> -ENOSPC; + space_info = __find_space_info(extent_root->fs_info, >> flags); if (!space_info) { ret = >> update_space_info(extent_root->fs_info, flags, @@ -3606,6 +3610,8 >> @@ again: goto again; } >> >> + trans->allocating_chunk = true; + /* * If we have mixed >> data/metadata chunks we want to make sure we keep * allocating >> mixed chunks instead of individual chunks. @@ -3632,6 +3638,7 @@ >> again: check_system_chunk(trans, extent_root, flags); >> >> ret = btrfs_alloc_chunk(trans, extent_root, flags); + >> trans->allocating_chunk = false; if (ret < 0 && ret != -ENOSPC) >> goto out; >> >> diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c >> index e6509b9..47ad8be 100644 --- a/fs/btrfs/transaction.c +++ >> b/fs/btrfs/transaction.c @@ -388,6 +388,7 @@ again: >> h->qgroup_reserved = qgroup_reserved; h->delayed_ref_elem.seq = >> 0; h->type = type; + h->allocating_chunk = false; >> INIT_LIST_HEAD(&h->qgroup_ref_list); >> INIT_LIST_HEAD(&h->new_bgs); >> >> diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h >> index 0e8aa1e..69700f7 100644 --- a/fs/btrfs/transaction.h +++ >> b/fs/btrfs/transaction.h @@ -68,6 +68,7 @@ struct >> btrfs_trans_handle { struct btrfs_block_rsv *orig_rsv; short >> aborted; short adding_csums; + bool allocating_chunk; enum >> btrfs_trans_type type; /* * this root is only needed to validate >> that the root passed to > > I hit this problem in a following scenario: - a data chunk > allocation is triggered, and locks chunk_mutex - the same thread > now also wants to allocate a metadata chunk, so it recursively > calls do_chunk_alloc, but cannot lock the chunk_mutex => deadlock - > btrfs has only one metadata chunk, the one that was initially > allocated by mkfs, it has: total_bytes=8388608 bytes_used=8130560 > bytes_pinned=77824 bytes_reserved=180224 so bytes_used + > bytes_pinned + bytes_reserved == total_bytes > > Your patch would have returned ENOSPC and avoid the deadlock, but > there would be a failure to allocate a tree block for metadata. So > the transaction would have probably aborted. > > How such situation should be handled? > > Idea1: - lock chunk mutex, - if we are allocating a data chunk, > check whether the metadata space is below some threshold. If yes, > go and allocate a metadata chunk first and then only a data chunk. > > Idea2: - check if we are the same thread that already locked the > chunk mutex. If yes, allow recursive call but don't attempt to > lock/unlock the chunk_mutex this time > > Or some other way? > I fixed this with the delayed chunk allocation stuff which doesn't actually do the block group creation stuff until we end the transaction, so we can allocate metadata chunks without any issue. Thanks, Josef -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJTA4UMAAoJEANb+wAKly3B+KEP/RdlEyJWydetjQxllF0cgHY1 UraqWBl+mSSHlwZlHyGjmAu6cK6n+QfTZtdIBhihdY50UcvMuWtVmz2JzlbxeO5+ 88dBevADmW+QQoRl0yyQgnjlLWm+LvMTgOd1r+DZqlGs6sdX05dMI207+fQOW+c4 P+UKbT/eUYRVC4K//J1GUk4Yh3Q70U25321RWCehSUciwDVJO2LztD9VBAgh3qUc o5uh5syshS3RbEi0hnUQ8tDKXWvdZQBA2RF4loXACCmQO95e84mxVpoYPd9S1yYs J+wf+Bak5hKZxmXJkOVcjLj4GsVQFJWTBTj6FvOFrm5TAFEGSyzrEzL8xW361+VS I1q8GPSVN1fGKkVypddylLIXLHmqXb57UElvGhoBM0otxNd8+xfSpLZ045vv5qLx RKwhJI1gIWD59kBre0fdSkUJZDeYSmLWOiwG6hG3A7Yy93c6/1RLHRnHq5NEe12R nrqZKBnkvDKnL/21eVqpOMo7i/AzCB7N+ojfaql2WvWcLkCpomhLBgC18Q1RiSzZ nfmafQIUPunM4l/fLXsbYFdiU
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
Hello Josef, On Tue, Dec 18, 2012 at 3:52 PM, Josef Bacik wrote: > On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: >> An user reported that he has hit an annoying deadlock while playing with >> ceph based on btrfs. >> >> Current updating device tree requires space from METADATA chunk, >> so we -may- need to do a recursive chunk allocation when adding/updating >> dev extent, that is where the deadlock comes from. >> >> If we use SYSTEM metadata to update device tree, we can avoid the recursive >> stuff. >> > > This is going to cause us to allocate much more system chunks than we used to > which could land us in trouble. Instead let's just keep us from re-entering > if > we're already allocating a chunk. We do the chunk allocation when we don't > have > enough space for a cluster, but we'll likely have plenty of space to make an > allocation. Can you give this patch a try Jim and see if it fixes your > problem? > Thanks, > > Josef > > > diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c > index e152809..59df5e7 100644 > --- a/fs/btrfs/extent-tree.c > +++ b/fs/btrfs/extent-tree.c > @@ -3564,6 +3564,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle > *trans, > int wait_for_alloc = 0; > int ret = 0; > > + /* Don't re-enter if we're already allocating a chunk */ > + if (trans->allocating_chunk) > + return -ENOSPC; > + > space_info = __find_space_info(extent_root->fs_info, flags); > if (!space_info) { > ret = update_space_info(extent_root->fs_info, flags, > @@ -3606,6 +3610,8 @@ again: > goto again; > } > > + trans->allocating_chunk = true; > + > /* > * If we have mixed data/metadata chunks we want to make sure we keep > * allocating mixed chunks instead of individual chunks. > @@ -3632,6 +3638,7 @@ again: > check_system_chunk(trans, extent_root, flags); > > ret = btrfs_alloc_chunk(trans, extent_root, flags); > + trans->allocating_chunk = false; > if (ret < 0 && ret != -ENOSPC) > goto out; > > diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c > index e6509b9..47ad8be 100644 > --- a/fs/btrfs/transaction.c > +++ b/fs/btrfs/transaction.c > @@ -388,6 +388,7 @@ again: > h->qgroup_reserved = qgroup_reserved; > h->delayed_ref_elem.seq = 0; > h->type = type; > + h->allocating_chunk = false; > INIT_LIST_HEAD(&h->qgroup_ref_list); > INIT_LIST_HEAD(&h->new_bgs); > > diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h > index 0e8aa1e..69700f7 100644 > --- a/fs/btrfs/transaction.h > +++ b/fs/btrfs/transaction.h > @@ -68,6 +68,7 @@ struct btrfs_trans_handle { > struct btrfs_block_rsv *orig_rsv; > short aborted; > short adding_csums; > + bool allocating_chunk; > enum btrfs_trans_type type; > /* > * this root is only needed to validate that the root passed to I hit this problem in a following scenario: - a data chunk allocation is triggered, and locks chunk_mutex - the same thread now also wants to allocate a metadata chunk, so it recursively calls do_chunk_alloc, but cannot lock the chunk_mutex => deadlock - btrfs has only one metadata chunk, the one that was initially allocated by mkfs, it has: total_bytes=8388608 bytes_used=8130560 bytes_pinned=77824 bytes_reserved=180224 so bytes_used + bytes_pinned + bytes_reserved == total_bytes Your patch would have returned ENOSPC and avoid the deadlock, but there would be a failure to allocate a tree block for metadata. So the transaction would have probably aborted. How such situation should be handled? Idea1: - lock chunk mutex, - if we are allocating a data chunk, check whether the metadata space is below some threshold. If yes, go and allocate a metadata chunk first and then only a data chunk. Idea2: - check if we are the same thread that already locked the chunk mutex. If yes, allow recursive call but don't attempt to lock/unlock the chunk_mutex this time Or some other way? Thanks! Alex. > -- > To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/31/2013 08:33 AM, Josef Bacik wrote: > On Wed, Jan 30, 2013 at 02:37:40PM -0700, Jim Schutt wrote: >> On 01/30/2013 09:38 AM, Josef Bacik wrote: >>> On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > On 01/29/2013 01:04 PM, Josef Bacik wrote: >>> On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: >>> On 01/28/2013 02:23 PM, Josef Bacik wrote: >>> On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: >>> Hi Josef, >>> >>> Thanks for the patch - sorry for the long delay in >>> testing... >>> >>> >>> Jim, >>> >>> I've been trying to reason out how this happens, could you do a >>> btrfs fi df on >>> the filesystem thats giving you trouble so I can see if what I >>> think is >>> happening is what's actually happening. Thanks, >>> >>> Here's an example, using a slightly different kernel than >>> my previous report. It's your btrfs-next master branch >>> (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") >>> with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). >>> >>> >>> Here I'm finding the file system in question: >>> >>> # ls -l /dev/mapper | grep dm-93 >>> lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 >>> >>> # df -h | grep -A 1 cs53s19p2 >>> /dev/mapper/cs53s19p2 >>> 896G 1.1G 896G 1% >>> /ram/mnt/ceph/data.osd.522 >>> >>> >>> Here's the info you asked for: >>> >>> # btrfs fi df /ram/mnt/ceph/data.osd.522 >>> Data: total=2.01GB, used=1.00GB >>> System: total=4.00MB, used=64.00KB >>> Metadata: total=8.00MB, used=7.56MB >>> >>> How big is the disk you are using, and what mount options? I have a >>> patch to >>> keep the panic from happening and hopefully the abort, could you try >>> this? I >>> still want to keep the underlying error from happening because it >>> shouldn't be, >>> but no reason I can't fix the error case while you can easily reproduce >>> it :). >>> Thanks, >>> >>> Josef >>> >>> >From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 >>> From: Josef Bacik >>> Date: Tue, 29 Jan 2013 15:03:37 -0500 >>> Subject: [PATCH] Btrfs: fix chunk allocation error handling >>> >>> If we error out allocating a dev extent we will have already created the >>> block group and such which will cause problems since the allocator may >>> have >>> tried to allocate out of the block group that no longer exists. This >>> will >>> cause BUG_ON()'s in the bio submission path. This also makes a failure >>> to >>> allocate a dev extent a non-abort error, we will just clean up the dev >>> extents we did allocate and exit. Now if we fail to delete the dev >>> extents >>> we will abort since we can't have half of the dev extents hanging >>> around, >>> but this will make us much less likely to abort. Thanks, >>> >>> Signed-off-by: Josef Bacik >>> --- > > Interesting - with your patch applied I triggered the following, just > bringing up a fresh Ceph filesystem - I didn't even get a chance to > mount it on my Ceph clients: > >>> Ok can you give this patch a whirl as well? It seems to fix the problem >>> for me. >> >> With this patch on top of your previous patch, after several trials of >> my test I am also unable to reproduce the issue. Since I had been >> having trouble first time, every time, I think it also seems to fix >> the problem for me. >> > > Hey Jim, > > Could you test this patch instead? I think it's a little less hamfisted and > should give us a nice balance between not crashing and being good for > performance. Thanks, Hi Josef, Running with this patch in place of your previous version, I was again unable to reproduce the issue. I might be seeing a couple percent increase in performance, or it might just be noise, but I'm willing to say that I think performance is same-or-better than the previous version of the patch. Thanks again! -- Jim > > Josef > > commit 43510c0e5faad8e5e4d8ba13baa1dd5dfb3d39ce > Author: Josef Bacik > Date: Wed Jan 30 17:02:51 2013 -0500 > > Btrfs: do not allow overcommit to happen if we are over 80% in use > > Because of how little we allocate chunks now we can get really tight on > metadata space before we will allocate a new chunk. This resulted in > being > unable to add device extents when allocating a new metadata chunk as we > did > not have enough space. This is because we were allowed to overcommit too > much metadata without actually making sure we
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Wed, Jan 30, 2013 at 02:37:40PM -0700, Jim Schutt wrote: > On 01/30/2013 09:38 AM, Josef Bacik wrote: > > On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > >> > On 01/29/2013 01:04 PM, Josef Bacik wrote: > >>> > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > > > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: > >>> > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > > > >> Hi Josef, > > > > >> > > > > >> Thanks for the patch - sorry for the long delay in > > > > >> testing... > > > > >> > >>> > >>> > > > >>> > >>> > > Jim, > >>> > >>> > > > >>> > >>> > > I've been trying to reason out how this happens, could you > >>> > >>> > > do a btrfs fi df on > >>> > >>> > > the filesystem thats giving you trouble so I can see if > >>> > >>> > > what I think is > >>> > >>> > > happening is what's actually happening. Thanks, > > > >> > > > > >> > Here's an example, using a slightly different kernel than > > > >> > my previous report. It's your btrfs-next master branch > > > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > > > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > > > >> > > > > >> > > > > >> > Here I'm finding the file system in question: > > > >> > > > > >> > # ls -l /dev/mapper | grep dm-93 > > > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > > > >> > > > > >> > # df -h | grep -A 1 cs53s19p2 > > > >> > /dev/mapper/cs53s19p2 > > > >> > 896G 1.1G 896G 1% > > > >> > /ram/mnt/ceph/data.osd.522 > > > >> > > > > >> > > > > >> > Here's the info you asked for: > > > >> > > > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > > > >> > Data: total=2.01GB, used=1.00GB > > > >> > System: total=4.00MB, used=64.00KB > > > >> > Metadata: total=8.00MB, used=7.56MB > > > >> > > >>> > > How big is the disk you are using, and what mount options? I have a > >>> > > patch to > >>> > > keep the panic from happening and hopefully the abort, could you try > >>> > > this? I > >>> > > still want to keep the underlying error from happening because it > >>> > > shouldn't be, > >>> > > but no reason I can't fix the error case while you can easily > >>> > > reproduce it :). > >>> > > Thanks, > >>> > > > >>> > > Josef > >>> > > > >>> > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > >>> > > From: Josef Bacik > >>> > > Date: Tue, 29 Jan 2013 15:03:37 -0500 > >>> > > Subject: [PATCH] Btrfs: fix chunk allocation error handling > >>> > > > >>> > > If we error out allocating a dev extent we will have already created > >>> > > the > >>> > > block group and such which will cause problems since the allocator > >>> > > may have > >>> > > tried to allocate out of the block group that no longer exists. This > >>> > > will > >>> > > cause BUG_ON()'s in the bio submission path. This also makes a > >>> > > failure to > >>> > > allocate a dev extent a non-abort error, we will just clean up the dev > >>> > > extents we did allocate and exit. Now if we fail to delete the dev > >>> > > extents > >>> > > we will abort since we can't have half of the dev extents hanging > >>> > > around, > >>> > > but this will make us much less likely to abort. Thanks, > >>> > > > >>> > > Signed-off-by: Josef Bacik > >>> > > --- > >> > > >> > Interesting - with your patch applied I triggered the following, just > >> > bringing up a fresh Ceph filesystem - I didn't even get a chance to > >> > mount it on my Ceph clients: > >> > > > Ok can you give this patch a whirl as well? It seems to fix the problem > > for me. > > With this patch on top of your previous patch, after several trials of > my test I am also unable to reproduce the issue. Since I had been > having trouble first time, every time, I think it also seems to fix > the problem for me. > Hey Jim, Could you test this patch instead? I think it's a little less hamfisted and should give us a nice balance between not crashing and being good for performance. Thanks, Josef commit 43510c0e5faad8e5e4d8ba13baa1dd5dfb3d39ce Author: Josef Bacik Date: Wed Jan 30 17:02:51 2013 -0500 Btrfs: do not allow overcommit to happen if we are over 80% in use Because of how little we allocate chunks now we can get really tight on metadata space before we will allocate a new chunk. This resulted in being unable to add device extents when allocating a new metadata chunk as we did not have enough space. This is because we were allowed to overcommit too much metadata without actually making sure we had enough space to make allocations. The idea behind overcommit is that we are allowed to say "sure you can have that reservation" when most of the free space is occupied by rese
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Wed, Jan 30, 2013 at 02:37:40PM -0700, Jim Schutt wrote: > On 01/30/2013 09:38 AM, Josef Bacik wrote: > > On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > >> > On 01/29/2013 01:04 PM, Josef Bacik wrote: > >>> > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > > > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: > >>> > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > > > >> Hi Josef, > > > > >> > > > > >> Thanks for the patch - sorry for the long delay in > > > > >> testing... > > > > >> > >>> > >>> > > > >>> > >>> > > Jim, > >>> > >>> > > > >>> > >>> > > I've been trying to reason out how this happens, could you > >>> > >>> > > do a btrfs fi df on > >>> > >>> > > the filesystem thats giving you trouble so I can see if > >>> > >>> > > what I think is > >>> > >>> > > happening is what's actually happening. Thanks, > > > >> > > > > >> > Here's an example, using a slightly different kernel than > > > >> > my previous report. It's your btrfs-next master branch > > > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > > > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > > > >> > > > > >> > > > > >> > Here I'm finding the file system in question: > > > >> > > > > >> > # ls -l /dev/mapper | grep dm-93 > > > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > > > >> > > > > >> > # df -h | grep -A 1 cs53s19p2 > > > >> > /dev/mapper/cs53s19p2 > > > >> > 896G 1.1G 896G 1% > > > >> > /ram/mnt/ceph/data.osd.522 > > > >> > > > > >> > > > > >> > Here's the info you asked for: > > > >> > > > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > > > >> > Data: total=2.01GB, used=1.00GB > > > >> > System: total=4.00MB, used=64.00KB > > > >> > Metadata: total=8.00MB, used=7.56MB > > > >> > > >>> > > How big is the disk you are using, and what mount options? I have a > >>> > > patch to > >>> > > keep the panic from happening and hopefully the abort, could you try > >>> > > this? I > >>> > > still want to keep the underlying error from happening because it > >>> > > shouldn't be, > >>> > > but no reason I can't fix the error case while you can easily > >>> > > reproduce it :). > >>> > > Thanks, > >>> > > > >>> > > Josef > >>> > > > >>> > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > >>> > > From: Josef Bacik > >>> > > Date: Tue, 29 Jan 2013 15:03:37 -0500 > >>> > > Subject: [PATCH] Btrfs: fix chunk allocation error handling > >>> > > > >>> > > If we error out allocating a dev extent we will have already created > >>> > > the > >>> > > block group and such which will cause problems since the allocator > >>> > > may have > >>> > > tried to allocate out of the block group that no longer exists. This > >>> > > will > >>> > > cause BUG_ON()'s in the bio submission path. This also makes a > >>> > > failure to > >>> > > allocate a dev extent a non-abort error, we will just clean up the dev > >>> > > extents we did allocate and exit. Now if we fail to delete the dev > >>> > > extents > >>> > > we will abort since we can't have half of the dev extents hanging > >>> > > around, > >>> > > but this will make us much less likely to abort. Thanks, > >>> > > > >>> > > Signed-off-by: Josef Bacik > >>> > > --- > >> > > >> > Interesting - with your patch applied I triggered the following, just > >> > bringing up a fresh Ceph filesystem - I didn't even get a chance to > >> > mount it on my Ceph clients: > >> > > > Ok can you give this patch a whirl as well? It seems to fix the problem > > for me. > > With this patch on top of your previous patch, after several trials of > my test I am also unable to reproduce the issue. Since I had been > having trouble first time, every time, I think it also seems to fix > the problem for me. > > Thanks again! > Awesome thanks for testing! Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/30/2013 09:38 AM, Josef Bacik wrote: > On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: >> > On 01/29/2013 01:04 PM, Josef Bacik wrote: >>> > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: >>> > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > > >> Hi Josef, > > > >> > > > >> Thanks for the patch - sorry for the long delay in > > > >> testing... > > > >> >>> > >>> > > >>> > >>> > > Jim, >>> > >>> > > >>> > >>> > > I've been trying to reason out how this happens, could you do >>> > >>> > > a btrfs fi df on >>> > >>> > > the filesystem thats giving you trouble so I can see if what >>> > >>> > > I think is >>> > >>> > > happening is what's actually happening. Thanks, > > >> > > > >> > Here's an example, using a slightly different kernel than > > >> > my previous report. It's your btrfs-next master branch > > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > > >> > > > >> > > > >> > Here I'm finding the file system in question: > > >> > > > >> > # ls -l /dev/mapper | grep dm-93 > > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > > >> > > > >> > # df -h | grep -A 1 cs53s19p2 > > >> > /dev/mapper/cs53s19p2 > > >> > 896G 1.1G 896G 1% > > >> > /ram/mnt/ceph/data.osd.522 > > >> > > > >> > > > >> > Here's the info you asked for: > > >> > > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > > >> > Data: total=2.01GB, used=1.00GB > > >> > System: total=4.00MB, used=64.00KB > > >> > Metadata: total=8.00MB, used=7.56MB > > >> > >>> > > How big is the disk you are using, and what mount options? I have a >>> > > patch to >>> > > keep the panic from happening and hopefully the abort, could you try >>> > > this? I >>> > > still want to keep the underlying error from happening because it >>> > > shouldn't be, >>> > > but no reason I can't fix the error case while you can easily reproduce >>> > > it :). >>> > > Thanks, >>> > > >>> > > Josef >>> > > >>> > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 >>> > > From: Josef Bacik >>> > > Date: Tue, 29 Jan 2013 15:03:37 -0500 >>> > > Subject: [PATCH] Btrfs: fix chunk allocation error handling >>> > > >>> > > If we error out allocating a dev extent we will have already created the >>> > > block group and such which will cause problems since the allocator may >>> > > have >>> > > tried to allocate out of the block group that no longer exists. This >>> > > will >>> > > cause BUG_ON()'s in the bio submission path. This also makes a failure >>> > > to >>> > > allocate a dev extent a non-abort error, we will just clean up the dev >>> > > extents we did allocate and exit. Now if we fail to delete the dev >>> > > extents >>> > > we will abort since we can't have half of the dev extents hanging >>> > > around, >>> > > but this will make us much less likely to abort. Thanks, >>> > > >>> > > Signed-off-by: Josef Bacik >>> > > --- >> > >> > Interesting - with your patch applied I triggered the following, just >> > bringing up a fresh Ceph filesystem - I didn't even get a chance to >> > mount it on my Ceph clients: >> > > Ok can you give this patch a whirl as well? It seems to fix the problem for > me. With this patch on top of your previous patch, after several trials of my test I am also unable to reproduce the issue. Since I had been having trouble first time, every time, I think it also seems to fix the problem for me. Thanks again! -- Jim > Thanks, > > Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > On 01/29/2013 01:04 PM, Josef Bacik wrote: > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > >> Hi Josef, > > >> > > >> Thanks for the patch - sorry for the long delay in testing... > > >> > >>> > > > >>> > > Jim, > >>> > > > >>> > > I've been trying to reason out how this happens, could you do a btrfs > >>> > > fi df on > >>> > > the filesystem thats giving you trouble so I can see if what I think > >>> > > is > >>> > > happening is what's actually happening. Thanks, > >> > > >> > Here's an example, using a slightly different kernel than > >> > my previous report. It's your btrfs-next master branch > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > >> > > >> > > >> > Here I'm finding the file system in question: > >> > > >> > # ls -l /dev/mapper | grep dm-93 > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > >> > > >> > # df -h | grep -A 1 cs53s19p2 > >> > /dev/mapper/cs53s19p2 > >> > 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 > >> > > >> > > >> > Here's the info you asked for: > >> > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > >> > Data: total=2.01GB, used=1.00GB > >> > System: total=4.00MB, used=64.00KB > >> > Metadata: total=8.00MB, used=7.56MB > >> > > > How big is the disk you are using, and what mount options? I have a patch > > to > > keep the panic from happening and hopefully the abort, could you try this? > > I > > still want to keep the underlying error from happening because it shouldn't > > be, > > but no reason I can't fix the error case while you can easily reproduce it > > :). > > Thanks, > > > > Josef > > > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > > From: Josef Bacik > > Date: Tue, 29 Jan 2013 15:03:37 -0500 > > Subject: [PATCH] Btrfs: fix chunk allocation error handling > > > > If we error out allocating a dev extent we will have already created the > > block group and such which will cause problems since the allocator may have > > tried to allocate out of the block group that no longer exists. This will > > cause BUG_ON()'s in the bio submission path. This also makes a failure to > > allocate a dev extent a non-abort error, we will just clean up the dev > > extents we did allocate and exit. Now if we fail to delete the dev extents > > we will abort since we can't have half of the dev extents hanging around, > > but this will make us much less likely to abort. Thanks, > > > > Signed-off-by: Josef Bacik > > --- > > Interesting - with your patch applied I triggered the following, just > bringing up a fresh Ceph filesystem - I didn't even get a chance to > mount it on my Ceph clients: > Ok can you give this patch a whirl as well? It seems to fix the problem for me. Thanks, Josef diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index dca5679..874bcf2 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3677,8 +3677,18 @@ static int can_overcommit(struct btrfs_root *root, u64 used; used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; + space_info->bytes_pinned + space_info->bytes_readonly; + + /* +* We only want to allow over committing if we have lots of actual space +* free, but if we've tied up more than 80% of the space with actual +* space reservation (not including bytes we _might_ use) then don't +* allow overcommitting as it will just make things go badly for us. +*/ + if (used > div_factor(space_info->total_bytes, 8)) + return 0; + + used += space_info->bytes_may_use; spin_lock(&root->fs_info->free_chunk_lock); avail = root->fs_info->free_chunk_space; -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > On 01/29/2013 01:04 PM, Josef Bacik wrote: > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > >> Hi Josef, > > >> > > >> Thanks for the patch - sorry for the long delay in testing... > > >> > >>> > > > >>> > > Jim, > >>> > > > >>> > > I've been trying to reason out how this happens, could you do a btrfs > >>> > > fi df on > >>> > > the filesystem thats giving you trouble so I can see if what I think > >>> > > is > >>> > > happening is what's actually happening. Thanks, > >> > > >> > Here's an example, using a slightly different kernel than > >> > my previous report. It's your btrfs-next master branch > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > >> > > >> > > >> > Here I'm finding the file system in question: > >> > > >> > # ls -l /dev/mapper | grep dm-93 > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > >> > > >> > # df -h | grep -A 1 cs53s19p2 > >> > /dev/mapper/cs53s19p2 > >> > 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 > >> > > >> > > >> > Here's the info you asked for: > >> > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > >> > Data: total=2.01GB, used=1.00GB > >> > System: total=4.00MB, used=64.00KB > >> > Metadata: total=8.00MB, used=7.56MB > >> > > > How big is the disk you are using, and what mount options? I have a patch > > to > > keep the panic from happening and hopefully the abort, could you try this? > > I > > still want to keep the underlying error from happening because it shouldn't > > be, > > but no reason I can't fix the error case while you can easily reproduce it > > :). > > Thanks, > > > > Josef > > > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > > From: Josef Bacik > > Date: Tue, 29 Jan 2013 15:03:37 -0500 > > Subject: [PATCH] Btrfs: fix chunk allocation error handling > > > > If we error out allocating a dev extent we will have already created the > > block group and such which will cause problems since the allocator may have > > tried to allocate out of the block group that no longer exists. This will > > cause BUG_ON()'s in the bio submission path. This also makes a failure to > > allocate a dev extent a non-abort error, we will just clean up the dev > > extents we did allocate and exit. Now if we fail to delete the dev extents > > we will abort since we can't have half of the dev extents hanging around, > > but this will make us much less likely to abort. Thanks, > > > > Signed-off-by: Josef Bacik > > --- > > Interesting - with your patch applied I triggered the following, just > bringing up a fresh Ceph filesystem - I didn't even get a chance to > mount it on my Ceph clients: > Actually nevermind it looks like I figured out how to reproduce. I'll let you know when I have something to test. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 04:05:17PM -0700, Jim Schutt wrote: > On 01/29/2013 01:04 PM, Josef Bacik wrote: > > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: > >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > >> Hi Josef, > > >> > > >> Thanks for the patch - sorry for the long delay in testing... > > >> > >>> > > > >>> > > Jim, > >>> > > > >>> > > I've been trying to reason out how this happens, could you do a btrfs > >>> > > fi df on > >>> > > the filesystem thats giving you trouble so I can see if what I think > >>> > > is > >>> > > happening is what's actually happening. Thanks, > >> > > >> > Here's an example, using a slightly different kernel than > >> > my previous report. It's your btrfs-next master branch > >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > >> > > >> > > >> > Here I'm finding the file system in question: > >> > > >> > # ls -l /dev/mapper | grep dm-93 > >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > >> > > >> > # df -h | grep -A 1 cs53s19p2 > >> > /dev/mapper/cs53s19p2 > >> > 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 > >> > > >> > > >> > Here's the info you asked for: > >> > > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 > >> > Data: total=2.01GB, used=1.00GB > >> > System: total=4.00MB, used=64.00KB > >> > Metadata: total=8.00MB, used=7.56MB > >> > > > How big is the disk you are using, and what mount options? I have a patch > > to > > keep the panic from happening and hopefully the abort, could you try this? > > I > > still want to keep the underlying error from happening because it shouldn't > > be, > > but no reason I can't fix the error case while you can easily reproduce it > > :). > > Thanks, > > > > Josef > > > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > > From: Josef Bacik > > Date: Tue, 29 Jan 2013 15:03:37 -0500 > > Subject: [PATCH] Btrfs: fix chunk allocation error handling > > > > If we error out allocating a dev extent we will have already created the > > block group and such which will cause problems since the allocator may have > > tried to allocate out of the block group that no longer exists. This will > > cause BUG_ON()'s in the bio submission path. This also makes a failure to > > allocate a dev extent a non-abort error, we will just clean up the dev > > extents we did allocate and exit. Now if we fail to delete the dev extents > > we will abort since we can't have half of the dev extents hanging around, > > but this will make us much less likely to abort. Thanks, > > > > Signed-off-by: Josef Bacik > > --- > > Interesting - with your patch applied I triggered the following, just > bringing up a fresh Ceph filesystem - I didn't even get a chance to > mount it on my Ceph clients: > Well that makes me a sad panda, but hey it didn't panic this time. What workload are you running on this fs/ceph cluster? Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/29/2013 01:04 PM, Josef Bacik wrote: > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: >> > On 01/28/2013 02:23 PM, Josef Bacik wrote: >>> > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > >> Hi Josef, > >> > >> Thanks for the patch - sorry for the long delay in testing... > >> >>> > > >>> > > Jim, >>> > > >>> > > I've been trying to reason out how this happens, could you do a btrfs >>> > > fi df on >>> > > the filesystem thats giving you trouble so I can see if what I think is >>> > > happening is what's actually happening. Thanks, >> > >> > Here's an example, using a slightly different kernel than >> > my previous report. It's your btrfs-next master branch >> > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") >> > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). >> > >> > >> > Here I'm finding the file system in question: >> > >> > # ls -l /dev/mapper | grep dm-93 >> > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 >> > >> > # df -h | grep -A 1 cs53s19p2 >> > /dev/mapper/cs53s19p2 >> > 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 >> > >> > >> > Here's the info you asked for: >> > >> > # btrfs fi df /ram/mnt/ceph/data.osd.522 >> > Data: total=2.01GB, used=1.00GB >> > System: total=4.00MB, used=64.00KB >> > Metadata: total=8.00MB, used=7.56MB >> > > How big is the disk you are using, and what mount options? I have a patch to > keep the panic from happening and hopefully the abort, could you try this? I > still want to keep the underlying error from happening because it shouldn't > be, > but no reason I can't fix the error case while you can easily reproduce it :). > Thanks, > > Josef > >>From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 > From: Josef Bacik > Date: Tue, 29 Jan 2013 15:03:37 -0500 > Subject: [PATCH] Btrfs: fix chunk allocation error handling > > If we error out allocating a dev extent we will have already created the > block group and such which will cause problems since the allocator may have > tried to allocate out of the block group that no longer exists. This will > cause BUG_ON()'s in the bio submission path. This also makes a failure to > allocate a dev extent a non-abort error, we will just clean up the dev > extents we did allocate and exit. Now if we fail to delete the dev extents > we will abort since we can't have half of the dev extents hanging around, > but this will make us much less likely to abort. Thanks, > > Signed-off-by: Josef Bacik > --- Interesting - with your patch applied I triggered the following, just bringing up a fresh Ceph filesystem - I didn't even get a chance to mount it on my Ceph clients: [ 6419.450179] BTRFS error (device dm-73) in btrfs_free_dev_extent:1115: error 28 (Slot search failed) [ 6419.459223] btrfs is forced readonly [ 6419.462805] [ cut here ] [ 6419.467440] WARNING: at fs/btrfs/super.c:256 __btrfs_abort_transaction+0x60/0x110 [btrfs]() [ 6419.475809] Hardware name: X8DTH-i/6/iF/6F [ 6419.479914] Modules linked in: btrfs zlib_deflate ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ib_addr ipv6 dm_mirror dm_region_hash dm_log dm_round_robin dm_multipath scsi_dh vhost_net macvtap macvlan tun uinput sg joydev sd_mod iTCO_wdt iTCO_vendor_support hid_generic coretemp kvm crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode button ata_piix libata mpt2sas scsi_transport_sas raid_class scsi_mod serio_raw pcspkr mlx4_ib ib_sa ib_mad ib_core mlx4_en mlx4_core cxgb4 i2c_i801 i2c_core lpc_ich mfd_core uhci_hcd ehci_hcd i7core_edac edac_core ioatdma dm_mod nfsv4 auth_rpcgss nfsv3 nfs_acl nfsv2 nfs lockd sunrpc fscache broadcom tg3 hwmon bnx2 igb dca e1000 [ 6419.546095] Pid: 107593, comm: ceph-osd Not tainted 3.7.0-00270-g8353482 #494 [ 6419.553227] Call Trace: [ 6419.555697] [] warn_slowpath_common+0x94/0xc0 [ 6419.561708] [] warn_slowpath_fmt+0x46/0x50 [ 6419.567491] [] __btrfs_abort_transaction+0x60/0x110 [btrfs] [ 6419.574746] [] __btrfs_alloc_chunk+0x6e6/0x770 [btrfs] [ 6419.581553] [] btrfs_alloc_chunk+0x5e/0x90 [btrfs] [ 6419.588017] [] ? check_system_chunk+0x71/0x130 [btrfs] [ 6419.594824] [] do_chunk_alloc+0x2ec/0x370 [btrfs] [ 6419.601188] [] find_free_extent+0xaac/0xbe0 [btrfs] [ 6419.607733] [] btrfs_reserve_extent+0x82/0x190 [btrfs] [ 6419.614545] [] btrfs_alloc_free_block+0x85/0x230 [btrfs] [ 6419.621530] [] ? check_buffer_tree_ref+0x25/0x50 [btrfs] [ 6419.628512] [] __btrfs_cow_block+0x14a/0x4b0 [btrfs] [ 6419.635155] [] ? btrfs_try_tree_write_lock+0x3c/0xa0 [btrfs] [ 6419.642475] [] ? btrfs_set_lock_blocking_rw+0xe3/0x160 [btrfs] [ 6419.649970] [] btrfs_cow_block+0x161/0x200 [btrfs] [ 6419.656424] [] btrfs_search_slot+0x399/0x760 [btrfs] [ 6419.663050] [] btrfs_truncate_inode_items+0x179/0x710 [btrfs] [ 6419.670458] [] ? btrfs_add_ordered_operation+0x5
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/29/2013 01:04 PM, Josef Bacik wrote: > On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: >> On 01/28/2013 02:23 PM, Josef Bacik wrote: >>> On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: Hi Josef, Thanks for the patch - sorry for the long delay in testing... >>> >>> Jim, >>> >>> I've been trying to reason out how this happens, could you do a btrfs fi df >>> on >>> the filesystem thats giving you trouble so I can see if what I think is >>> happening is what's actually happening. Thanks, >> >> Here's an example, using a slightly different kernel than >> my previous report. It's your btrfs-next master branch >> (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") >> with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). >> >> >> Here I'm finding the file system in question: >> >> # ls -l /dev/mapper | grep dm-93 >> lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 >> >> # df -h | grep -A 1 cs53s19p2 >> /dev/mapper/cs53s19p2 >> 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 >> >> >> Here's the info you asked for: >> >> # btrfs fi df /ram/mnt/ceph/data.osd.522 >> Data: total=2.01GB, used=1.00GB >> System: total=4.00MB, used=64.00KB >> Metadata: total=8.00MB, used=7.56MB >> > > How big is the disk you are using, and what mount options? The partition is ~900 GiB, and the mount options according to /proc/mount are: rw,noatime,nospace_cache Also, in case it matters, I build the file systems with -l 65536 -n 65536. > I have a patch to > keep the panic from happening and hopefully the abort, could you try this? I > still want to keep the underlying error from happening because it shouldn't > be, > but no reason I can't fix the error case while you can easily reproduce it :). I'm happy to try it - but I probably won't have results for you until tomorrow, due to other time pressures. Thanks for taking a look. -- Jim > Thanks, > > Josef > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 11:41:10AM -0700, Jim Schutt wrote: > On 01/28/2013 02:23 PM, Josef Bacik wrote: > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > >> Hi Josef, > >> > >> Thanks for the patch - sorry for the long delay in testing... > >> > > > > Jim, > > > > I've been trying to reason out how this happens, could you do a btrfs fi df > > on > > the filesystem thats giving you trouble so I can see if what I think is > > happening is what's actually happening. Thanks, > > Here's an example, using a slightly different kernel than > my previous report. It's your btrfs-next master branch > (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") > with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). > > > Here I'm finding the file system in question: > > # ls -l /dev/mapper | grep dm-93 > lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 > > # df -h | grep -A 1 cs53s19p2 > /dev/mapper/cs53s19p2 > 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 > > > Here's the info you asked for: > > # btrfs fi df /ram/mnt/ceph/data.osd.522 > Data: total=2.01GB, used=1.00GB > System: total=4.00MB, used=64.00KB > Metadata: total=8.00MB, used=7.56MB > How big is the disk you are using, and what mount options? I have a patch to keep the panic from happening and hopefully the abort, could you try this? I still want to keep the underlying error from happening because it shouldn't be, but no reason I can't fix the error case while you can easily reproduce it :). Thanks, Josef >From c50b725c74c7d39064e553ef85ac9753efbd8aec Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 29 Jan 2013 15:03:37 -0500 Subject: [PATCH] Btrfs: fix chunk allocation error handling If we error out allocating a dev extent we will have already created the block group and such which will cause problems since the allocator may have tried to allocate out of the block group that no longer exists. This will cause BUG_ON()'s in the bio submission path. This also makes a failure to allocate a dev extent a non-abort error, we will just clean up the dev extents we did allocate and exit. Now if we fail to delete the dev extents we will abort since we can't have half of the dev extents hanging around, but this will make us much less likely to abort. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/volumes.c | 32 ++-- 1 files changed, 22 insertions(+), 10 deletions(-) diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 4f8c281..2ba5b84 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3766,12 +3766,6 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, if (ret) goto error; - ret = btrfs_make_block_group(trans, extent_root, 0, type, -BTRFS_FIRST_CHUNK_TREE_OBJECTID, -start, num_bytes); - if (ret) - goto error; - for (i = 0; i < map->num_stripes; ++i) { struct btrfs_device *device; u64 dev_offset; @@ -3783,15 +3777,33 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, info->chunk_root->root_key.objectid, BTRFS_FIRST_CHUNK_TREE_OBJECTID, start, dev_offset, stripe_size); - if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); - goto error; - } + if (ret) + goto error_dev_extent; + } + + ret = btrfs_make_block_group(trans, extent_root, 0, type, +BTRFS_FIRST_CHUNK_TREE_OBJECTID, +start, num_bytes); + if (ret) { + i = map->num_stripes - 1; + goto error_dev_extent; } kfree(devices_info); return 0; +error_dev_extent: + for (; i >= 0; i--) { + struct btrfs_device *device; + int err; + + device = map->stripes[i].dev; + err = btrfs_free_dev_extent(trans, device, start); + if (err) { + btrfs_abort_transaction(trans, extent_root, err); + break; + } + } error: kfree(map); kfree(devices_info); -- 1.7.7.6 -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/28/2013 02:23 PM, Josef Bacik wrote: > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: >> Hi Josef, >> >> Thanks for the patch - sorry for the long delay in testing... >> > > Jim, > > I've been trying to reason out how this happens, could you do a btrfs fi df on > the filesystem thats giving you trouble so I can see if what I think is > happening is what's actually happening. Thanks, Here's an example, using a slightly different kernel than my previous report. It's your btrfs-next master branch (commit 8f139e59d5 "Btrfs: use bit operation for ->fs_state") with ceph 3.8 for-linus (commit 0fa6ebc600 from linus' tree). Here I'm finding the file system in question: # ls -l /dev/mapper | grep dm-93 lrwxrwxrwx 1 root root 8 Jan 29 11:13 cs53s19p2 -> ../dm-93 # df -h | grep -A 1 cs53s19p2 /dev/mapper/cs53s19p2 896G 1.1G 896G 1% /ram/mnt/ceph/data.osd.522 Here's the info you asked for: # btrfs fi df /ram/mnt/ceph/data.osd.522 Data: total=2.01GB, used=1.00GB System: total=4.00MB, used=64.00KB Metadata: total=8.00MB, used=7.56MB And here's the backtrace that had trouble on dm-93. It's a little different to my previous report: [ 705.496463] [ cut here ] [ 705.501123] WARNING: at fs/btrfs/super.c:256 __btrfs_abort_transaction+0x60/0x110 [btrfs]() [ 705.509751] Hardware name: X8DTH-i/6/iF/6F [ 705.513862] Modules linked in: btrfs zlib_deflate ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ib_addr ipv6 dm_mirror dm_region_hash dm_log dm_round_robin dm_multipath scsi_dh vhost_net macvtap macvlan tun uinput sg joydev sd_mod hid_generic iTCO_wdt iTCO_vendor_support coretemp kvm crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode serio_raw pcspkr mlx4_ib ib_sa ib_mad ib_core mlx4_en mlx4_core ata_piix libata mpt2sas scsi_transport_sas raid_class scsi_mod cxgb4 i2c_i801 i2c_core button lpc_ich mfd_core ehci_hcd uhci_hcd i7core_edac edac_core dm_mod ioatdma nfsv4 auth_rpcgss nfsv3 nfs_acl nfsv2 nfs lockd sunrpc fscache broadcom tg3 hwmon bnx2 igb dca e1000 [ 705.580232] Pid: 33025, comm: ceph-osd Not tainted 3.7.0-00269-gd9acbfd #492 [ 705.587488] Call Trace: [ 705.589957] [] warn_slowpath_common+0x94/0xc0 [ 705.596108] [] ? btrfs_free_path+0x2a/0x40 [btrfs] [ 705.602685] [] warn_slowpath_fmt+0x46/0x50 [ 705.608563] [] __btrfs_abort_transaction+0x60/0x110 [btrfs] [ 705.615994] [] __btrfs_alloc_chunk+0x678/0x710 [btrfs] [ 705.622945] [] btrfs_alloc_chunk+0x5e/0x90 [btrfs] [ 705.629635] [] ? check_system_chunk+0x71/0x130 [btrfs] [ 705.637079] [] do_chunk_alloc+0x2ec/0x370 [btrfs] [ 705.643451] [] ? btrfs_reduce_alloc_profile+0xa9/0x120 [btrfs] [ 705.650951] [] btrfs_check_data_free_space+0x13c/0x2b0 [btrfs] [ 705.658446] [] btrfs_delalloc_reserve_space+0x20/0x60 [btrfs] [ 705.665882] [] __btrfs_buffered_write+0x15e/0x340 [btrfs] [ 705.672952] [] btrfs_file_aio_write+0x309/0x450 [btrfs] [ 705.679889] [] ? __btrfs_direct_write+0x130/0x130 [btrfs] [ 705.686934] [] do_sync_readv_writev+0x94/0xe0 [ 705.692942] [] do_readv_writev+0xe3/0x1e0 [ 705.698604] [] ? fget_light+0x122/0x170 [ 705.704093] [] vfs_writev+0x46/0x60 [ 705.709239] [] sys_writev+0x5f/0xc0 [ 705.714388] [] ? trace_hardirqs_on_thunk+0x3a/0x3f [ 705.720827] [] system_call_fastpath+0x16/0x1b [ 705.726829] ---[ end trace 6e889d6d939ca116 ]--- [ 705.731459] BTRFS warning (device dm-93): __btrfs_alloc_chunk:3787: Aborting unused transaction(error 28). [ 705.741187] btrfs: mapping failed logical 1099431936 bio len 524288 len 65536 [ 705.741192] BTRFS warning (device dm-93): find_free_extent:5948: Aborting unused transaction(Object already exists). [ 705.759185] [ cut here ] [ 705.763929] kernel BUG at fs/btrfs/volumes.c:4891! [ 705.768990] invalid opcode: [#1] SMP [ 705.773561] Modules linked in: btrfs zlib_deflate ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ib_addr ipv6 dm_mirror dm_region_hash dm_log dm_round_robin dm_multipath scsi_dh vhost_net macvtap macvlan tun uinput sg joydev sd_mod hid_generic iTCO_wdt iTCO_vendor_support coretemp kvm crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode serio_raw pcspkr mlx4_ib ib_sa ib_mad ib_core mlx4_en mlx4_core ata_piix libata mpt2sas scsi_transport_sas raid_class scsi_mod cxgb4 i2c_i801 i2c_core button lpc_ich mfd_core ehci_hcd uhci_hcd i7core_edac edac_core dm_mod ioatdma nfsv4 auth_rpcgss nfsv3 nfs_acl nfsv2 nfs lockd sunrpc fscache broadcom tg3 hwmon bnx2 igb dca e1000 [ 705.845121] CPU 22 [ 705.847114] Pid: 21317, comm: btrfs-worker-1 Tainted: GW 3.7.0-00269-gd9acbfd #492 Supermicro X8DTH-i/6/iF/6F/X8DTH [ 705.858886] RIP: 0010:[] [] btrfs_map_bio+0x8d/0x300 [btrfs] [ 705.867928] RSP: 0018:880610ce7c58 EFLAGS: 00010296 [ 705.873363] RAX: 0041 RBX: 88061c
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 05:43:31PM +0100, David Sterba wrote: > On Tue, Jan 29, 2013 at 08:50:34AM -0500, Josef Bacik wrote: > You can use scsi_debug device with > > parm: lbpu:enable LBP, support UNMAP command (def=0) (int) Also, loop device with a file backed by a filesystem with hole punch support also understands TRIM. david -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 08:50:34AM -0500, Josef Bacik wrote: > On Tue, Jan 29, 2013 at 08:47:30AM -0500, Josef Bacik wrote: > > 251 [not run] FSTRIM is not supported > > > > Are you sure its 251? Thanks, > > Sorry it's early, I need a device that does trim. /me waits for his fusion > card > to get back from the shop, You can use scsi_debug device with parm: lbpu:enable LBP, support UNMAP command (def=0) (int) david -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Jan 29, 2013 at 08:47:30AM -0500, Josef Bacik wrote: > On Mon, Jan 28, 2013 at 07:30:09PM -0700, Liu Bo wrote: > > On Mon, Jan 28, 2013 at 04:23:31PM -0500, Josef Bacik wrote: > > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > > > Hi Josef, > > > > > > > > Thanks for the patch - sorry for the long delay in testing... > > > > > > > > > > Jim, > > > > > > I've been trying to reason out how this happens, could you do a btrfs fi > > > df on > > > the filesystem thats giving you trouble so I can see if what I think is > > > happening is what's actually happening. Thanks, > > > > Josef, > > > > A quick reproducer here: running xfstests 251 with autodefrag,compress=zlib > > > > > 251 [not run] FSTRIM is not supported > > Are you sure its 251? Thanks, Sorry it's early, I need a device that does trim. /me waits for his fusion card to get back from the shop, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Mon, Jan 28, 2013 at 07:30:09PM -0700, Liu Bo wrote: > On Mon, Jan 28, 2013 at 04:23:31PM -0500, Josef Bacik wrote: > > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > > Hi Josef, > > > > > > Thanks for the patch - sorry for the long delay in testing... > > > > > > > Jim, > > > > I've been trying to reason out how this happens, could you do a btrfs fi df > > on > > the filesystem thats giving you trouble so I can see if what I think is > > happening is what's actually happening. Thanks, > > Josef, > > A quick reproducer here: running xfstests 251 with autodefrag,compress=zlib > 251 [not run] FSTRIM is not supported Are you sure its 251? Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Mon, Jan 28, 2013 at 04:23:31PM -0500, Josef Bacik wrote: > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > > Hi Josef, > > > > Thanks for the patch - sorry for the long delay in testing... > > > > Jim, > > I've been trying to reason out how this happens, could you do a btrfs fi df on > the filesystem thats giving you trouble so I can see if what I think is > happening is what's actually happening. Thanks, Josef, A quick reproducer here: running xfstests 251 with autodefrag,compress=zlib thanks, liubo -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On 01/28/2013 02:23 PM, Josef Bacik wrote: > On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: >> Hi Josef, >> >> Thanks for the patch - sorry for the long delay in testing... >> > > Jim, > > I've been trying to reason out how this happens, could you do a btrfs fi df on > the filesystem thats giving you trouble so I can see if what I think is > happening is what's actually happening. Thanks, Sure - it'll take me a bit to set the test up again. -- Jim > > Josef > > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Thu, Jan 03, 2013 at 11:44:46AM -0700, Jim Schutt wrote: > Hi Josef, > > Thanks for the patch - sorry for the long delay in testing... > Jim, I've been trying to reason out how this happens, could you do a btrfs fi df on the filesystem thats giving you trouble so I can see if what I think is happening is what's actually happening. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
Hi Josef, Thanks for the patch - sorry for the long delay in testing... On 12/18/2012 06:52 AM, Josef Bacik wrote: > On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: >> An user reported that he has hit an annoying deadlock while playing with >> ceph based on btrfs. >> >> Current updating device tree requires space from METADATA chunk, >> so we -may- need to do a recursive chunk allocation when adding/updating >> dev extent, that is where the deadlock comes from. >> >> If we use SYSTEM metadata to update device tree, we can avoid the recursive >> stuff. >> > > This is going to cause us to allocate much more system chunks than we used to > which could land us in trouble. Instead let's just keep us from re-entering > if > we're already allocating a chunk. We do the chunk allocation when we don't > have > enough space for a cluster, but we'll likely have plenty of space to make an > allocation. Can you give this patch a try Jim and see if it fixes your > problem? > Thanks, > > Josef > With your patch applied to 3.7.1, I get the following on one of my servers running Ceph OSDs. The end effect is that some of my ceph client writes hang. [ 1440.335752] [ cut here ] [ 1440.340602] WARNING: at fs/btrfs/super.c:246 __btrfs_abort_transaction+0x60/0x110 [btrfs]() [ 1440.349117] Hardware name: X8DTH-i/6/iF/6F [ 1440.353252] Modules linked in: btrfs zlib_deflate ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ib_addr ipv6 dm_mirror dm_region_hash dm_log dm_round_robin dm_multipath scsi_dh vhost_net macvtap macvlan tun uinput sg joydev sd_mod iTCO_wdt iTCO_vendor_support hid_generic button ata_piix libata coretemp kvm crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode mpt2sas scsi_transport_sas raid_class scsi_mod serio_raw pcspkr mlx4_ib ib_sa ib_mad ib_core mlx4_en mlx4_core cxgb4 i2c_i801 i2c_core lpc_ich mfd_core ehci_hcd uhci_hcd ioatdma i7core_edac dm_mod edac_core nfsv4 auth_rpcgss nfsv3 nfs_acl nfsv2 nfs lockd sunrpc fscache broadcom tg3 hwmon bnx2 igb dca e1000 [ 1440.419398] Pid: 48686, comm: ceph-osd Not tainted 3.7.1-6-gc794580 #484 [ 1440.426614] Call Trace: [ 1440.429083] [] warn_slowpath_common+0x94/0xc0 [ 1440.435110] [] warn_slowpath_fmt+0x46/0x50 [ 1440.440894] [] __btrfs_abort_transaction+0x60/0x110 [btrfs] [ 1440.448135] [] __btrfs_alloc_chunk+0x6cd/0x750 [btrfs] [ 1440.454941] [] btrfs_alloc_chunk+0x5e/0x90 [btrfs] [ 1440.461382] [] ? check_system_chunk+0x71/0x130 [btrfs] [ 1440.468188] [] do_chunk_alloc+0x2ec/0x370 [btrfs] [ 1440.474562] [] ? btrfs_reduce_alloc_profile+0xa9/0x120 [btrfs] [ 1440.482050] [] btrfs_check_data_free_space+0x13c/0x2b0 [btrfs] [ 1440.489558] [] btrfs_delalloc_reserve_space+0x20/0x60 [btrfs] [ 1440.497013] [] __btrfs_buffered_write+0x15e/0x350 [btrfs] [ 1440.504095] [] btrfs_file_aio_write+0x209/0x320 [btrfs] [ 1440.511000] [] ? __btrfs_direct_write+0x130/0x130 [btrfs] [ 1440.518062] [] do_sync_readv_writev+0x94/0xe0 [ 1440.524105] [] do_readv_writev+0xe3/0x1e0 [ 1440.529792] [] ? fget_light+0x122/0x170 [ 1440.535275] [] vfs_writev+0x46/0x60 [ 1440.540412] [] sys_writev+0x5f/0xc0 [ 1440.545547] [] ? trace_hardirqs_on_thunk+0x3a/0x3f [ 1440.551987] [] system_call_fastpath+0x16/0x1b [ 1440.558016] ---[ end trace 764e83a458dabca6 ]--- [ 1440.562662] BTRFS warning (device dm-32): __btrfs_alloc_chunk:3488: Aborting unused transaction(error 28). [ 1440.595987] BTRFS warning (device dm-32): find_free_extent:5871: Aborting unused transaction(Object already exists). [ 1440.606542] BUG: unable to handle kernel NULL pointer dereference at (null) [ 1440.614382] IP: [] map_private_extent_buffer+0xe/0xf0 [btrfs] [ 1440.621704] PGD 6138e8067 PUD 56749f067 PMD 0 [ 1440.626190] Oops: [#1] SMP [ 1440.629442] Modules linked in: btrfs zlib_deflate ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ib_addr ipv6 dm_mirror dm_region_hash dm_log dm_round_robin dm_multipath scsi_dh vhost_net macvtap macvlan tun uinput sg joydev sd_mod iTCO_wdt iTCO_vendor_support hid_generic button ata_piix libata coretemp kvm crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 xts gf128mul microcode mpt2sas scsi_transport_sas raid_class scsi_mod serio_raw pcspkr mlx4_ib ib_sa ib_mad ib_core mlx4_en mlx4_core cxgb4 i2c_i801 i2c_core lpc_ich mfd_core ehci_hcd uhci_hcd ioatdma i7core_edac dm_mod edac_core nfsv4 auth_rpcgss nfsv3 nfs_acl nfsv2 nfs lockd sunrpc fscache broadcom tg3 hwmon bnx2 igb dca e1000 [ 1440.694855] CPU 16 [ 1440.696784] Pid: 48687, comm: ceph-osd Tainted: GW 3.7.1-6-gc794580 #484 Supermicro X8DTH-i/6/iF/6F/X8DTH [ 1440.707803] RIP: 0010:[] [] map_private_extent_buffer+0xe/0xf0 [btrfs] [ 1440.717544] RSP: 0018:880b740db9f8 EFLAGS: 00010292 [ 1440.722841] RAX: RBX: RCX: 880b740dba28 [ 1440.729947] RDX: 0004 RSI
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Dec 18, 2012 at 07:47:51AM -0700, Liu Bo wrote: > On Tue, Dec 18, 2012 at 08:52:42AM -0500, Josef Bacik wrote: > > On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: > > > An user reported that he has hit an annoying deadlock while playing with > > > ceph based on btrfs. > > > > > > Current updating device tree requires space from METADATA chunk, > > > so we -may- need to do a recursive chunk allocation when adding/updating > > > dev extent, that is where the deadlock comes from. > > > > > > If we use SYSTEM metadata to update device tree, we can avoid the > > > recursive > > > stuff. > > > > > > > This is going to cause us to allocate much more system chunks than we used > > to > > which could land us in trouble. Instead let's just keep us from > > re-entering if > > we're already allocating a chunk. We do the chunk allocation when we don't > > have > > enough space for a cluster, but we'll likely have plenty of space to make an > > allocation. Can you give this patch a try Jim and see if it fixes your > > problem? > > Thanks, > > From the stack info Jim gave, returning ENOSPC to caller will end up with > aborting to readonly if there is no others save the situation by > allocating another METADATA chunk, it is recursive allocation though. > if (ret < 0 && ret != -ENOSPC) it shouldn't abort, it should just drop empty_size and stop trying to allocate a cluster and just allocate the blocks needed, and this is only for the recursive chunk allocation, so after this succeeds we'll have a new chunk and the original allocation will be able to carry on. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Tue, Dec 18, 2012 at 08:52:42AM -0500, Josef Bacik wrote: > On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: > > An user reported that he has hit an annoying deadlock while playing with > > ceph based on btrfs. > > > > Current updating device tree requires space from METADATA chunk, > > so we -may- need to do a recursive chunk allocation when adding/updating > > dev extent, that is where the deadlock comes from. > > > > If we use SYSTEM metadata to update device tree, we can avoid the recursive > > stuff. > > > > This is going to cause us to allocate much more system chunks than we used to > which could land us in trouble. Instead let's just keep us from re-entering > if > we're already allocating a chunk. We do the chunk allocation when we don't > have > enough space for a cluster, but we'll likely have plenty of space to make an > allocation. Can you give this patch a try Jim and see if it fixes your > problem? > Thanks, >From the stack info Jim gave, returning ENOSPC to caller will end up with aborting to readonly if there is no others save the situation by allocating another METADATA chunk, it is recursive allocation though. thanks, liubo > > Josef > > > diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c > index e152809..59df5e7 100644 > --- a/fs/btrfs/extent-tree.c > +++ b/fs/btrfs/extent-tree.c > @@ -3564,6 +3564,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle > *trans, > int wait_for_alloc = 0; > int ret = 0; > > + /* Don't re-enter if we're already allocating a chunk */ > + if (trans->allocating_chunk) > + return -ENOSPC; > + > space_info = __find_space_info(extent_root->fs_info, flags); > if (!space_info) { > ret = update_space_info(extent_root->fs_info, flags, > @@ -3606,6 +3610,8 @@ again: > goto again; > } > > + trans->allocating_chunk = true; > + > /* >* If we have mixed data/metadata chunks we want to make sure we keep >* allocating mixed chunks instead of individual chunks. > @@ -3632,6 +3638,7 @@ again: > check_system_chunk(trans, extent_root, flags); > > ret = btrfs_alloc_chunk(trans, extent_root, flags); > + trans->allocating_chunk = false; > if (ret < 0 && ret != -ENOSPC) > goto out; > > diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c > index e6509b9..47ad8be 100644 > --- a/fs/btrfs/transaction.c > +++ b/fs/btrfs/transaction.c > @@ -388,6 +388,7 @@ again: > h->qgroup_reserved = qgroup_reserved; > h->delayed_ref_elem.seq = 0; > h->type = type; > + h->allocating_chunk = false; > INIT_LIST_HEAD(&h->qgroup_ref_list); > INIT_LIST_HEAD(&h->new_bgs); > > diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h > index 0e8aa1e..69700f7 100644 > --- a/fs/btrfs/transaction.h > +++ b/fs/btrfs/transaction.h > @@ -68,6 +68,7 @@ struct btrfs_trans_handle { > struct btrfs_block_rsv *orig_rsv; > short aborted; > short adding_csums; > + bool allocating_chunk; > enum btrfs_trans_type type; > /* >* this root is only needed to validate that the root passed to -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] Btrfs: fix a deadlock on chunk mutex
On Wed, Dec 12, 2012 at 06:52:37PM -0700, Liu Bo wrote: > An user reported that he has hit an annoying deadlock while playing with > ceph based on btrfs. > > Current updating device tree requires space from METADATA chunk, > so we -may- need to do a recursive chunk allocation when adding/updating > dev extent, that is where the deadlock comes from. > > If we use SYSTEM metadata to update device tree, we can avoid the recursive > stuff. > This is going to cause us to allocate much more system chunks than we used to which could land us in trouble. Instead let's just keep us from re-entering if we're already allocating a chunk. We do the chunk allocation when we don't have enough space for a cluster, but we'll likely have plenty of space to make an allocation. Can you give this patch a try Jim and see if it fixes your problem? Thanks, Josef diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index e152809..59df5e7 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3564,6 +3564,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, int wait_for_alloc = 0; int ret = 0; + /* Don't re-enter if we're already allocating a chunk */ + if (trans->allocating_chunk) + return -ENOSPC; + space_info = __find_space_info(extent_root->fs_info, flags); if (!space_info) { ret = update_space_info(extent_root->fs_info, flags, @@ -3606,6 +3610,8 @@ again: goto again; } + trans->allocating_chunk = true; + /* * If we have mixed data/metadata chunks we want to make sure we keep * allocating mixed chunks instead of individual chunks. @@ -3632,6 +3638,7 @@ again: check_system_chunk(trans, extent_root, flags); ret = btrfs_alloc_chunk(trans, extent_root, flags); + trans->allocating_chunk = false; if (ret < 0 && ret != -ENOSPC) goto out; diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index e6509b9..47ad8be 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -388,6 +388,7 @@ again: h->qgroup_reserved = qgroup_reserved; h->delayed_ref_elem.seq = 0; h->type = type; + h->allocating_chunk = false; INIT_LIST_HEAD(&h->qgroup_ref_list); INIT_LIST_HEAD(&h->new_bgs); diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index 0e8aa1e..69700f7 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -68,6 +68,7 @@ struct btrfs_trans_handle { struct btrfs_block_rsv *orig_rsv; short aborted; short adding_csums; + bool allocating_chunk; enum btrfs_trans_type type; /* * this root is only needed to validate that the root passed to -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PATCH] Btrfs: fix a deadlock on chunk mutex
An user reported that he has hit an annoying deadlock while playing with ceph based on btrfs. Current updating device tree requires space from METADATA chunk, so we -may- need to do a recursive chunk allocation when adding/updating dev extent, that is where the deadlock comes from. If we use SYSTEM metadata to update device tree, we can avoid the recursive stuff. Reported-by: Jim Schutt Signed-off-by: Liu Bo --- fs/btrfs/extent-tree.c |8 +--- 1 files changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 3d3e2c1..561dad5 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3346,7 +3346,8 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) if (data) flags = BTRFS_BLOCK_GROUP_DATA; - else if (root == root->fs_info->chunk_root) + else if (root == root->fs_info->chunk_root || +root == root->fs_info->dev_root) flags = BTRFS_BLOCK_GROUP_SYSTEM; else flags = BTRFS_BLOCK_GROUP_METADATA; @@ -3534,7 +3535,8 @@ static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type) else num_dev = 1;/* DUP or single */ - /* metadata for updaing devices and chunk tree */ + /* metadata for adding/updating devices and chunk tree */ + num_dev = num_dev << 1 return btrfs_calc_trans_metadata_size(root, num_dev + 1); } @@ -4351,7 +4353,7 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info) fs_info->extent_root->block_rsv = &fs_info->global_block_rsv; fs_info->csum_root->block_rsv = &fs_info->global_block_rsv; - fs_info->dev_root->block_rsv = &fs_info->global_block_rsv; + fs_info->dev_root->block_rsv = &fs_info->chunk_block_rsv; fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; -- 1.7.7.6 -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html