From pk@dbic.pro Sat Apr 25 12:35:27 2015 From: pk@dbic.pro To: openldap-bugs@openldap.org Subject: Re: (ITS#8089) LMDB hangs on mdb_txn_renew0 if MDB_NOSYNC flag is used Date: Sat, 25 Apr 2015 12:35:25 +0000 Message-ID: MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="===============6871322122835577567==" --===============6871322122835577567== Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable --90e6ba3fd5fbb4dffe05148bbe0e Content-Type: text/plain; charset=3DUTF-8 Hello Howard, Please accept my apologies, the issue was not related to LMDB. There have been a data race in my code that was only triggered by fast writes. Sorry for late answer, - your e-mail landed into the spam-folder, and I just found it. With best regards, Pavel. 2015-04-16 8:04 GMT+06:00 Howard Chu : > pk(a)dbic.pro wrote: > >> Full_Name: Pavel Kraynyukhov >> Version: only LMDB from git and lmdb-0.9.14 >> OS: Gentoo Linux >> URL: ftp://ftp.openldap.org/incoming/ >> Submission from: (NULL) (37.99.47.171) >> >> >> Hello there, I have ran into an issue with LMDB, which I can reproduce at >> will, >> but it seems a specific one. The git commit is >> 3368d1f5e243225cba4d730fba19ff600798ebe3 >> And this commit and my issue seems to be related. >> >> 1. I use MDB_NOTLS flag on environment open. >> 2. if I use MDB_NOSYNC flag for LMDB environment, the writer thread is >> stuck >> after several transactions (writer thread backtrace): >> > > P.S. tested against system provided lmdb version 0.9.14 and result is the >> same. >> So this maybe not related to latest commit in git. >> > > If you're having this problem in 0.9.14 then it is certainly not related > to 3368d1f5e243225cba4d730fba19ff600798ebe3. > > You confirm that the issue does not occur in 0.9.13? Can you upload a test > case demonstrating the problem? > > -- > -- Howard Chu > CTO, Symas Corp. http://www.symas.com > Director, Highland Sun http://highlandsun.com/hyc/ > Chief Architect, OpenLDAP http://www.openldap.org/project/ > --90e6ba3fd5fbb4dffe05148bbe0e Content-Type: text/html; charset=3DUTF-8 Content-Transfer-Encoding: quoted-printable
Hello Howard,
Please accept m= =3D y apologies, the issue was not related to LMDB. There have been a data race=3D in my code that was only triggered by fast writes.
Sorry for late=3D answer, - your e-mail landed into the spam-folder, and I just found it.
With best regards,
Pavel.

2015-04-16 8:04 GMT+0= 6:=3D 00 Howard Chu <hyc(a)symas.com>:
pk(a)dbic.pro wro= te:
Full_Name: Pavel Kraynyukhov
Version: only LMDB from git and lmdb-0.9.14
OS: Gentoo Linux
URL: ftp:/= /f=3D tp.openldap.org/incoming/
Submission from: (NULL) (37.99.47.171)


Hello there, I have ran into an issue with LMDB, which I can reproduce at w=3D ill,
but it seems a specific one. The git commit is
3368d1f5e243225cba4d730fba19ff600798ebe3
And this commit and my issue seems to be related.

1. I use MDB_NOTLS flag on environment open.
2. if=3DC2=3DA0 I use MDB_NOSYNC flag for LMDB environment, the writer thread= i=3D s stuck
after several transactions (writer thread backtrace):

P.S. tested against system provided lmdb version 0.9.14 and result is the s=3D ame.
So this maybe not related to latest commit in git.

If you're having this problem in 0.9.14 then it is certainly not relate=3D d to 3368d1f5e243225cba4d730fba19ff600798ebe3.

You confirm that the issue does not occur in 0.9.13? Can you upload a test =3D case demonstrating the problem?

--
=3DC2=3DA0 -- Howard Chu
=3DC2=3DA0 CTO, Symas Corp.=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2= =3DA0 =3DC2=3DA0http://www.symas.com
=3DC2=3DA0 Director, Highland Sun=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0http://highlandsun.com/hyc/
=3DC2=3DA0 Chief Architect, OpenLDAP=3DC2=3DA0 http://www.openldap.org/project/


--90e6ba3fd5fbb4dffe05148bbe0e-- --===============6871322122835577567==--