openldap-bugs

openldap-bugs@openldap.org

1 participants
20965 discussions

Re: (ITS#8976) LMDB > use posix_madvise(3) instead of madvise(3) for SmartOS
by matthieu.guegan＠smile-suisse.com 15 Feb '19

15 Feb '19

On Fri, Feb 15, 2019 at 4:33 PM Howard Chu <hyc(a)symas.com> wrote: > > Matthieu GUEGAN wrote: > > On Fri, Feb 15, 2019 at 3:56 PM Howard Chu <hyc(a)symas.com> wrote: > >> > >> mguegan(a)virtua.ch wrote: > >>> Full_Name: Matthieu Guegan > >>> Version: current > >>> OS: SmartOS > >>> URL: > >>> Submission from: (NULL) (109.190.148.77) > >>> > >>> > >>> Hi, > >>> > >>> In order to compile knot[1] on SmartOS, I have done a series of little patches. > >>> One of them is linked to the LMDB project, which redirects me here. > >>> > >>> The idea is to tell SmartOS (and I think the SunOS family) to make use of > >>> posix_madvise(3), instead of madvise(3) as implementation on this OS[2] is > >>> different than, for example, the Linux[3] one. > >>> > >>> So, I kindly ask you to take a look at the merge request of the knot project[4], > >>> in particular the `src/contrib/lmdb/mdb.c` patch (that's just a ifdef > >>> modification) : would it be possible to apply this patch directly in this > >>> repository ? > >> > >> I see no functional difference between your references [2] and [3]. > >> What is the benefit of this patch? > > > > Well, the Solaris/SunOS's madvise(3) is defined with an old `caddr_t` > > type whereas Linux (and *BSD OSes too) use a more modern approach with > > `void *`. > > On the `knot` project, which embed the lmdb part of openldap, > > compilation failed when using madvise(3) on SmartOS, so I did a > > replacement with posix_madvise(3) which uses the `void *` argument > > too. > > I don't know what's the best way to handle this case, that's why I ask here. > > A caddr_t is just a typedef of "char *". It should always be legal to > pass a void * to such a parameter. What is your compile error message? > ``` contrib/lmdb/mdb.c: In function 'mdb_env_map': contrib/lmdb/mdb.c:3998:3: error: implicit declaration of function 'madvise'; did you mean 'raise'? [-Werror=implicit-function-declaration] madvise(env->me_map, env->me_mapsize, MADV_RANDOM); ^~~~~~~ raise ```

1 0

Re: (ITS#8976) LMDB > use posix_madvise(3) instead of madvise(3) for SmartOS
by hyc＠symas.com 15 Feb '19

15 Feb '19

Matthieu GUEGAN wrote: > On Fri, Feb 15, 2019 at 3:56 PM Howard Chu <hyc(a)symas.com> wrote: >> >> mguegan(a)virtua.ch wrote: >>> Full_Name: Matthieu Guegan >>> Version: current >>> OS: SmartOS >>> URL: >>> Submission from: (NULL) (109.190.148.77) >>> >>> >>> Hi, >>> >>> In order to compile knot[1] on SmartOS, I have done a series of little patches. >>> One of them is linked to the LMDB project, which redirects me here. >>> >>> The idea is to tell SmartOS (and I think the SunOS family) to make use of >>> posix_madvise(3), instead of madvise(3) as implementation on this OS[2] is >>> different than, for example, the Linux[3] one. >>> >>> So, I kindly ask you to take a look at the merge request of the knot project[4], >>> in particular the `src/contrib/lmdb/mdb.c` patch (that's just a ifdef >>> modification) : would it be possible to apply this patch directly in this >>> repository ? >> >> I see no functional difference between your references [2] and [3]. >> What is the benefit of this patch? > > Well, the Solaris/SunOS's madvise(3) is defined with an old `caddr_t` > type whereas Linux (and *BSD OSes too) use a more modern approach with > `void *`. > On the `knot` project, which embed the lmdb part of openldap, > compilation failed when using madvise(3) on SmartOS, so I did a > replacement with posix_madvise(3) which uses the `void *` argument > too. > I don't know what's the best way to handle this case, that's why I ask here. A caddr_t is just a typedef of "char *". It should always be legal to pass a void * to such a parameter. What is your compile error message? -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8976) LMDB > use posix_madvise(3) instead of madvise(3) for SmartOS
by matthieu.guegan＠smile-suisse.com 15 Feb '19

15 Feb '19

On Fri, Feb 15, 2019 at 3:56 PM Howard Chu <hyc(a)symas.com> wrote: > > mguegan(a)virtua.ch wrote: > > Full_Name: Matthieu Guegan > > Version: current > > OS: SmartOS > > URL: > > Submission from: (NULL) (109.190.148.77) > > > > > > Hi, > > > > In order to compile knot[1] on SmartOS, I have done a series of little patches. > > One of them is linked to the LMDB project, which redirects me here. > > > > The idea is to tell SmartOS (and I think the SunOS family) to make use of > > posix_madvise(3), instead of madvise(3) as implementation on this OS[2] is > > different than, for example, the Linux[3] one. > > > > So, I kindly ask you to take a look at the merge request of the knot project[4], > > in particular the `src/contrib/lmdb/mdb.c` patch (that's just a ifdef > > modification) : would it be possible to apply this patch directly in this > > repository ? > > I see no functional difference between your references [2] and [3]. > What is the benefit of this patch? Well, the Solaris/SunOS's madvise(3) is defined with an old `caddr_t` type whereas Linux (and *BSD OSes too) use a more modern approach with `void *`. On the `knot` project, which embed the lmdb part of openldap, compilation failed when using madvise(3) on SmartOS, so I did a replacement with posix_madvise(3) which uses the `void *` argument too. I don't know what's the best way to handle this case, that's why I ask here. > > > > Thanks > > > > > > [1] https://www.knot-dns.cz/ > > [2] https://docs.oracle.com/cd/E36784_01/html/E36874/madvise-3c.html > > [3] http://man7.org/linux/man-pages/man2/madvise.2.html > > [4] https://gitlab.labs.nic.cz/knot/knot-dns/merge_requests/979/diffs?commit_id… > > > > > > > -- > -- Howard Chu > CTO, Symas Corp. http://www.symas.com > Director, Highland Sun http://highlandsun.com/hyc/ > Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8976) LMDB > use posix_madvise(3) instead of madvise(3) for SmartOS
by hyc＠symas.com 15 Feb '19

15 Feb '19

mguegan(a)virtua.ch wrote: > Full_Name: Matthieu Guegan > Version: current > OS: SmartOS > URL: > Submission from: (NULL) (109.190.148.77) > > > Hi, > > In order to compile knot[1] on SmartOS, I have done a series of little patches. > One of them is linked to the LMDB project, which redirects me here. > > The idea is to tell SmartOS (and I think the SunOS family) to make use of > posix_madvise(3), instead of madvise(3) as implementation on this OS[2] is > different than, for example, the Linux[3] one. > > So, I kindly ask you to take a look at the merge request of the knot project[4], > in particular the `src/contrib/lmdb/mdb.c` patch (that's just a ifdef > modification) : would it be possible to apply this patch directly in this > repository ? I see no functional difference between your references [2] and [3]. What is the benefit of this patch? > > Thanks > > > [1] https://www.knot-dns.cz/ > [2] https://docs.oracle.com/cd/E36784_01/html/E36874/madvise-3c.html > [3] http://man7.org/linux/man-pages/man2/madvise.2.html > [4] https://gitlab.labs.nic.cz/knot/knot-dns/merge_requests/979/diffs?commit_id… > > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

(ITS#8977) Make IDL sizes configurable in back-mdb
by hyc＠openldap.org 15 Feb '19

15 Feb '19

Full_Name: Howard Chu Version: 2.5 OS: URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (80.233.42.124) Submitted by: hyc IDLs used for index slots are hardcoded to a max of 65536 elements in the DB and twice that in memory. Sites with larger DBs tend to need larger IDLs; if they're running with sufficient memory it ought to be possible to configure larger IDLs.

1 0

(ITS#8976) LMDB > use posix_madvise(3) instead of madvise(3) for SmartOS
by mguegan＠virtua.ch 15 Feb '19

15 Feb '19

Full_Name: Matthieu Guegan Version: current OS: SmartOS URL: Submission from: (NULL) (109.190.148.77) Hi, In order to compile knot[1] on SmartOS, I have done a series of little patches. One of them is linked to the LMDB project, which redirects me here. The idea is to tell SmartOS (and I think the SunOS family) to make use of posix_madvise(3), instead of madvise(3) as implementation on this OS[2] is different than, for example, the Linux[3] one. So, I kindly ask you to take a look at the merge request of the knot project[4], in particular the `src/contrib/lmdb/mdb.c` patch (that's just a ifdef modification) : would it be possible to apply this patch directly in this repository ? Thanks [1] https://www.knot-dns.cz/ [2] https://docs.oracle.com/cd/E36784_01/html/E36874/madvise-3c.html [3] http://man7.org/linux/man-pages/man2/madvise.2.html [4] https://gitlab.labs.nic.cz/knot/knot-dns/merge_requests/979/diffs?commit_id…

1 0

Re: (ITS#8841) Fix for an assertion error in back-meta
by nivanova＠symas.com 14 Feb '19

14 Feb '19

A new patch with added comments has been added at ftp://ftp.openldap.org/incoming/nadezhda-ivanova-190214.patch More details about this issue: The customer reported that when they simulated a "network unreachable" error, SLAPD exited with the following assertion error: #0 0x00007fa1d19e4885 in raise () from /lib64/libc.so.6 #1 0x00007fa1d19e5e61 in abort () from /lib64/libc.so.6 #2 0x00007fa1d19dd740 in __assert_fail () from /lib64/libc.so.6 #3 0x0000000000449c17 in slap_send_ldap_result (op=0x1870000, rs=0x7fa17c742930) at result.c:830 #4 0x00000000004b6a53 in meta_back_search (op=0x1870000, rs=0x7fa17c742930) at search.c:1183 #5 0x000000000043aef1 in fe_op_search (op=0x1870000, rs=0x7fa17c742930) at search.c:402 #6 0x00000000004a1a62 in overlay_op_walk (op=0x1870000, rs=0x7fa17c742930, which=op_search, oi=0x88583c0, on=0x0) at backover.c:677 #7 0x00000000004a24dc in over_op_func (op=0x1870000, rs=0x8, which=op_add) at backover.c:730 #8 0x000000000043b73c in do_search (op=0x1870000, rs=0x7fa17c742930) at search.c:247 #9 0x00000000004382af in connection_operation (ctx=0x7fa17c742b30, arg_v=<optimized out>) at connection.c:1316 #10 0x0000000000438edb in connection_read_thread (ctx=0x7fa17c742b30, argv=<optimized out>) at connection.c:1452 #11 0x00000000004f06a0 in ldap_int_thread_pool_wrapper (xpool=<optimized out>) at tpool.c:696 #12 0x00007fa1d1d35806 in start_thread () from /lib64/libpthread.so.0 #13 0x00007fa1d1a9065d in clone () from /lib64/libc.so.6 #14 0x0000000000000000 in ?? () This happened because meta used to directly return the ldap_sasl_bind return value to the client, and in this case it is not an LDAP error. We fixed the issue by mapping it to an LDAP error. Nadezhda Ivanova Software Engineer Symas Corporation http://www.symas.com

1 0

Re: (ITS#8847) New LDAP URL syntax to support binding to specific IP address at client side
by hyc＠symas.com 13 Feb '19

13 Feb '19

Sharma, Ramakant 2. (Nokia - IN/Bangalore) wrote: > Hi Howard, > > Please provide your valuable inputs for the below patch. It is a little strange that ldap_set_option() takes a string but ldap_get_option() returns a binary structure. get/set should be totally symmetric. It would be simplest to just save a copy of the string given to set and return the string in get. In libldap/os-ip.c, you should not be using ldap_get_option() to retrieve the values. You're operating inside libldap, you can access the structure directly. In particular, it's a waste to use ldap_get_option here and create a dynamically allocated copy of the address struct. Look at how the existing code works. Notice that it just uses ld->ld_options directly. This is rule #1 in software development - when extending existing code, your new code should do things the same way existing code does. Corollaries to this rule are: don't patch code without reading it first, and don't patch code without studying its complete context. > > Also let us know how to proceed further. > > BR, > Ramakant Sharma > > -----Original Message----- > From: Sharma, Ramakant 2. (Nokia - IN/Bangalore) > Sent: Monday, January 28, 2019 10:10 AM > To: Howard Chu <hyc(a)symas.com>; openldap-its(a)OpenLDAP.org > Cc: Singam, Sudhir (Nokia - IN/Bangalore) <sudhir.singam(a)nokia.com> > Subject: RE: (ITS#8847) New LDAP URL syntax to support binding to specific IP address at client side > > Hi Howard, > > I have uploaded the new patch which addresses the previous comments. > > Kindly check and provide your valuable comments . > > ftp://ftp.openldap.org/incoming/ramakant-sharma-190121.patch > > BR, > Ramakant Sharma > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8975) Calling mdb_env_set_mapsize on a shared DB fails sometimes
by hyc＠symas.com 10 Feb '19

10 Feb '19

kriszyp(a)gmail.com wrote: > Full_Name: Kristopher William Zyp > Version: LMDB 0.9.23 > OS: Windows Server 2012 R2, Windows 10 > URL: https://github.com/kriszyp/node-lmdb/commit/6df903907f5516320e9a8afce45bd32… > Submission from: (NULL) (71.199.6.148) > Thanks for the report and patch, added to mdb.RE/0.9 > Calling mdb_env_set_mapsize to increase the map size, when a DB is also in use > by other processes, when MDB_WRITEMAP is enabled (and the db file has been > opened with PAGE_READWRITE access), on Windows, will occasionally (seems like > about 1/100 attempts fails) produce an error "The requested operation cannot be > performed on a file with a user-mapped section open", or segfaults. The error > occurs in the SetFilePointer (or SetEndOfFile) call in mdb_env_map that is > performed to increase the allocated file size to the map size, prior to > CreateFileMapping. > > As it turns out this is pretty easy to solve, because manually expanding the > file size is not necessary when calling CreateFileMapping with PAGE_READWRITE > access, as Windows will automatically expand the file size for us, when opened > with the page write access enabled. Of course, this means all processes must be > consistent in use of MDB_WRITEMAP, but the documentation already makes this > explicit and clear. > > I believe this can be fixed by simply adding a check for MDB_WRITEMAP in the if > statement that calls SetFilePointer: > > if (!(flags & MDB_WRITEMAP) && (SetFilePointer(env->me_fd, sizelo, &sizehi, 0) > != (DWORD)sizelo > || !SetEndOfFile(env->me_fd) > || SetFilePointer(env->me_fd, 0, NULL, 0) != 0)) > return ErrCode(); > > The attached URL has the change as a patch/diff as applied to our node package. > > I am certainly happy to just keep this change on our own branches. There may be > nuances of this that I might not be aware of, but it seems to be working great > for us and I have tested this with MDB_WRITEMAP enabled and disabled. So I > thought I would offer/suggest this change, as it seems like it is > straightforward change to improve stability. Thank you! > > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

(ITS#8975) Calling mdb_env_set_mapsize on a shared DB fails sometimes
by kriszyp＠gmail.com 08 Feb '19

08 Feb '19

Full_Name: Kristopher William Zyp Version: LMDB 0.9.23 OS: Windows Server 2012 R2, Windows 10 URL: https://github.com/kriszyp/node-lmdb/commit/6df903907f5516320e9a8afce45bd32… Submission from: (NULL) (71.199.6.148) Calling mdb_env_set_mapsize to increase the map size, when a DB is also in use by other processes, when MDB_WRITEMAP is enabled (and the db file has been opened with PAGE_READWRITE access), on Windows, will occasionally (seems like about 1/100 attempts fails) produce an error "The requested operation cannot be performed on a file with a user-mapped section open", or segfaults. The error occurs in the SetFilePointer (or SetEndOfFile) call in mdb_env_map that is performed to increase the allocated file size to the map size, prior to CreateFileMapping. As it turns out this is pretty easy to solve, because manually expanding the file size is not necessary when calling CreateFileMapping with PAGE_READWRITE access, as Windows will automatically expand the file size for us, when opened with the page write access enabled. Of course, this means all processes must be consistent in use of MDB_WRITEMAP, but the documentation already makes this explicit and clear. I believe this can be fixed by simply adding a check for MDB_WRITEMAP in the if statement that calls SetFilePointer: if (!(flags & MDB_WRITEMAP) && (SetFilePointer(env->me_fd, sizelo, &sizehi, 0) != (DWORD)sizelo || !SetEndOfFile(env->me_fd) || SetFilePointer(env->me_fd, 0, NULL, 0) != 0)) return ErrCode(); The attached URL has the change as a patch/diff as applied to our node package. I am certainly happy to just keep this change on our own branches. There may be nuances of this that I might not be aware of, but it seems to be working great for us and I have tested this with MDB_WRITEMAP enabled and disabled. So I thought I would offer/suggest this change, as it seems like it is straightforward change to improve stability. Thank you!

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs