Re: Bugreport: openldap 2.4.48-1 ldapsearch coredump
--On Wednesday, September 4, 2019 2:47 PM +0300 Alexander Voropay wrote: Hi! Ildapsearch coredumps on (semi) complicated filter on MS AD LDAP (Filtered request to find all non-blocked users w/o e-mail address). The _same_ request works on the RHEL 7 ldapsearch "openldap-2.4.44-21.el7_6" Does it work on RHEL7 with OpenLDAP 2.4.48? You can get RHEL7 packages from <https://repo.symas.com/sofl/rhel7/> or <https://ltb-project.org/documentation/openldap-rpm#yum_repository> --Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Bug report: Killing a native process may not actually kill it
--On Wednesday, August 28, 2019 2:33 PM -0700 Kaz Kylheku <920-082-4...@kylheku.com> wrote: Cygwin can't introduce Unix-like shutdown mechanisms (like the handling a non-fatal signal) into non-Cygwin processes which have no concept of that. It makes no sense. My original post contained a link to a patch allowing for Cygwin to correctly terminate native Windows processes. I understand it is not the position of the Cygwin project to deal with situation, so I think we can just let it drop. Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Bug report: Killing a native process may not actually kill it
--On Wednesday, August 28, 2019 6:45 PM +0200 Corinna Vinschen wrote: Not likely. Cygwin handles Ctrl-C by generating SIGINT. This only works reliably with Cygwin processes. There's $ /bin/kill -f to call the Win32 function TerminateProcess(pid) on a non-Cygwin process or an unresponsive Cygwin process. As I noted, it was not unique to control-C. In any case, unfortunate to hear that Cygwin will not address this issue. kill -f is clearly not desirable for doing a clean shutdown of a process. --Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Bug report: Killing a native process may not actually kill it
--On Thursday, July 25, 2019 11:32 AM -0700 Quanah Gibson-Mount wrote: As found and reported to the MSYS team back in 2006 by Howard Chu, if a native process is spawned, control-C, the kill command, etc, may not actually kill the process. Details are here: I haven't seen a reply to this, so I just wanted to confirm that the Cygwin project is aware and (hopefully) will be able to do something to fix this in the long term. Thanks! Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: More on Win7 update messing up HOME
--On Saturday, August 17, 2019 12:28 AM +0300 Andrey Repin wrote: Greetings, David Karr! > I wish I could get my own messages on this list, so I could add more > information to my first note. Just reply to your own initial message. Or better yet, subscribe. I am subscribed. I can't reply to my own initial message because I never receive my own postings, I only receive notes sent by other people. You WROTE it. No need to receive to reply to it. As a subscriber, I receive a copy of every mail I send to the list (and I prefer it that way). Seems to be the default list setting. Perhaps David changed his list preferences at some point or there was a different default setting for the list at the time he subscribed? It's pretty common for this to be the default behavior for a mailing list. For one thing, it ensures you know that your email to the list was actually delivered. --Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Openldap 2.4.48-1 vs my company's pki
--On Monday, August 05, 2019 5:06 PM -0400 David Goldberg wrote: Correct, openssl s_client works, as does the older build of ldapsearch. I can't find any .ldaprc nor ldap.conf files on my system. Unfortunately I've only set up my system for end user purposes. Building from source will be a challenge. Any guidance (a link is fine) on what packages to install to set that up? And do I need to worry about the .cygport and patch files in the source distribution or will configure pick them up? I would start with executing ldapsearch with the -d -1 flag added in (full debugging) to see what the client is doing. I.e., if you're doing startTLS, then something like ldapsearch -x -ZZ -d -1 -H ldap:// -s base -b "" Or if you're using ldaps, then something like: ldapsearch -x -d -1 -H ldaps:// -s base -b "" Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Openldap 2.4.48-1 vs my company's pki
--On Monday, August 05, 2019 9:22 AM -0400 David Goldberg wrote: Sorry, was away from work over the weekend. I just tested with openssl s_client and it works just fine. Version is 1.1.1. there is no self signed certificate. It's signed with the company pki rather than commercial and I've properly installed that chain. The problem send to be with the new build, at least the weird ldd output leads me to that conclusion. I'll try to find some time to build from source and see if it Do you mean you connected to the ldap server using OpenSSL s_client to confirm that works? If that works and the ldapsearch (or other ldap client) binary does not, then you likely have a global /etc/ldap.conf (or whereever this build looks for it) or a ~/.ldaprc file that defines the path or file to find the CA certificate that would need updating. Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Openldap 2.4.48-1 vs my company's pki
--On Friday, August 02, 2019 12:45 PM -0400 David Goldberg wrote: I updated openldap from 2.4.42-1 to 2.4.48-1 this morning and now ldapsearch will not connect, complaining that the server provided certificate is self signed. I have set up /etc/pki with my company's certificate chain and that allows 2.4.42-1 (and earlier) and other applications to properly authenticate local services. What has changed in 2.4.48-1 that causes this to not work and how can I fix it. I've downgraded for now; that is not a good long term solution of course. What SSL library is being used for each of the two builds (I.e., gnutls? openssl? moznss?) What SSL library version did 2.4.42 link to? What SSL library version does 2.4.48 link to? Generally OpenLDAP should be linked to OpenSSL which uses PEM formatted certificates. Also check whether you have a global ldap.conf file (usually something like /etc/openldap/ldap.conf or /etc/ldap.conf, etc, depending on how OpenLDAP was built) that defines where to find the CA Cert(s), or a ~user/.ldaprc, etc. OpenLDAP client utilities generally by default do not search for a global list of CA certificates. --Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: openldap on Cygwin
--On Thursday, July 25, 2019 9:30 PM +0200 Achim Gratz wrote: Quanah Gibson-Mount writes: […] Sorry for wedging in sideways, but I've looked into building a more up-to-date openldap and there's missing detection / configuration for Cygwin. Specifically, there's code trying to use robust POSIX mutexes, which Cygwin doesn't have. As there is no configure option (that I could find), the solution is a bit invasive: either feeding in a number of preprocessor defines or patching in the recognition of Cygwin into libraries/liblmdb/mdb.c to define the correct options there. Any chance the upcoming release would have a fix for that? Hi Achim, I would suggest filing a report in the OpenLDAP issue tracker at: <https://www.openldap.org/its> I use MSYS2 for my OpenLDAP on windows builds, so haven't any direct experience at the moment with building it via cygwin. Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug report: Killing a native process may not actually kill it
As found and reported to the MSYS team back in 2006 by Howard Chu, if a native process is spawned, control-C, the kill command, etc, may not actually kill the process. Details are here: <http://mingw.5.n7.nabble.com/Re-Ctrl-Break-handler-td28010.html> as well as here: <https://sourceforge.net/p/mingw/bugs/1783/> Given that there is now 64-bit windows, the issue would be a bit more complex to resolve, as an updated patch would need to query first to see if the target process is 32-bit or 64-bit, and then a 64-bit version of the patch needs to be created. The inability to properly kill such a spawned process can lead to things like the dreaded "Device or resource busy" error, since the process that is using the related file has actually never terminated. It's fairly trivial to reproduce this in the OpenLDAP test suite, particularly test001, which spawns a slapd process, kills it, then spawns another slapd process, which will fail because the original slapd never actually got killed. Regards, Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com> -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Sunday, September 12, 2010 1:43 PM +0200 Corinna Vinschen wrote: On Sep 11 12:41, Corinna Vinschen wrote: On Sep 10 10:48, Quanah Gibson-Mount wrote: --On Friday, September 10, 2010 7:09 PM +0200 Corinna Vinschen wrote: Let me know if there is anything else I can provide. I'm not sure. I don't think so. The problem is that the unlink(2) function in Cygwin does not get any error code from any of the OS functions it calls. So, from the Cygwin POV everything worked fine. How is it supposed to know that anything has gone wrong, if the underlying OS doesn't tell? Heh, magic I guess. If I mount the drive as a CIFS drive from a Linux box, I can delete the files just fine, so for now that gives me a workaround (I'll move my deletion process to a Linux box). This morning I had an idea. While we were looking into the ACL, we neglected the DOS attributes. When you call `attrib' on one of the files for which you didn't call chmod yet, is the R/O attribute set? If so, it *could* explain why Cygwin thought it has successfully deleted the file, but it hasn't. I also might have a workaround for this. I've checked in a change which probably fixed your issue. The only exception are Cygwin symlinks of the old .lnk type, which has more than one link. That should occur rather seldom. Please test the next developer's snapshot from http://cygwin.com/snapshots/ Hi Corinna, I will give the snapshot a test. Here is the output of attrib: bu...@zre-win-002 /cygdrive/z/current/WINDOWS/ZDESKTOP-608/20100912050101_ZDESKTOP/ZimbraBuild/templates $ attrib BUILD_ISYNC_template AR Z:\current\WINDOWS\ZDESKTOP-608\20100912050101_ZDESKTOP\ZimbraBuild\templates\BUILD_ISYNC_template --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Sunday, September 12, 2010 2:21 PM -0700 Quanah Gibson-Mount wrote: Hi Corinna, I will give the snapshot a test. Here is the output of attrib: bu...@zre-win-002 /cygdrive/z/current/WINDOWS/ZDESKTOP-608/20100912050101_ZDESKTOP/ZimbraBu ild/templates $ attrib BUILD_ISYNC_template AR Z:\current\WINDOWS\ZDESKTOP-608\20100912050101_ZDESKTOP\ZimbraBuild\templ ates\BUILD_ISYNC_template Hi Corinna, With the snapshot in place, I can now remove the files. With plain rm, it prompts (as it should) and succeeds if I answer yes. With rm -f, no prompt, and it succeeds. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Friday, September 10, 2010 11:16 AM +0200 Corinna Vinschen wrote: On Sep 8 15:17, Quanah Gibson-Mount wrote: I have a CIFS drive I connect to as the windows user. I can write to the drive with no problem. However, when I go to delete files from the drive, Cygwin behaves very oddly. bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/tem plates $ rm -f * If you call rm w/o the -f flag, what error message do you get? Just a simple Permission denied, I guess. Nope. It doesn't give any error. bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ rm BUILD_ISYNC_template rm: remove write-protected regular file `BUILD_ISYNC_template'? y bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ ls -l BUILD_ISYNC_template -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_ISYNC_template bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/tem plates $ ls -l total 104 -r-xr-xr-x 1 1362 2010-09-08 13:31 BUILD_EVO_template -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_ISYNC_template [...] Now, if I modify the file to be +w, then -w, so it returns to its original permissions, I can suddenly delete it: Did you create the files with a Cygwin aplication or with a native Win32 application? In theory, there's nothing mysterious here, if the permissions of the file are so that the DELETE permission for your user or group is missing in the file's ACL. For obvious reasons the POSIX permission bits can't reflect the complexity of the original NT ACL. The chmod +w/-w somehow overwrite the original permissions with POSIX permissions as Cygwin generates them and the result is more DELETE friendly. Did you try to compare the ACL before and after the chmod? The output of `cacls filename' is probably different. The files are created with a native Win32 application (Perforce), where it is checking these files out of the Perforce repository. Here is the output from cacls prior to +w/-w: bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ cacls BUILD_ISYNC_template Z:\current\WINDOWS\main\20100908131458_ZDESKTOP\ZimbraBuild\templates\BUILD_ISYNC_template Everyone:F Account Domain not foundF Account Domain not foundR Everyone:R Here is cacls after +w/-w: Z:\current\WINDOWS\main\20100908131458_ZDESKTOP\ZimbraBuild\templates\BUILD_ISYNC_template Account Domain not found(special access:) STANDARD_RIGHTS_ALL DELETE READ_CONTROL WRITE_DAC WRITE_OWNER SYNCHRONIZE STANDARD_RIGHTS_REQUIRED FILE_GENERIC_READ FILE_GENERIC_EXECUTE FILE_READ_DATA FILE_READ_EA FILE_EXECUTE FILE_READ_ATTRIBUTES FILE_WRITE_ATTRIBUTES Account Domain not foundR Everyone:R This behavior is quite bizarre. I should be able to delete the files I created with the -f option to rm. Well, in theory, yes. However, it's possible that your CIFS drive has semantics which disallow this with the original ACL for some reason. Can you pleae run `strace -o rm.trace rm some_file' on a file which has the original ACL (before the chmod call) and send the rm.trace file? Done. I've provided strace output from both rm FILE and rm -f FILE Let me know if there is anything else I can provide. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration rm.trace.gz Description: Binary data rmf.trace.gz Description: Binary data -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Friday, September 10, 2010 7:09 PM +0200 Corinna Vinschen wrote: Let me know if there is anything else I can provide. I'm not sure. I don't think so. The problem is that the unlink(2) function in Cygwin does not get any error code from any of the OS functions it calls. So, from the Cygwin POV everything worked fine. How is it supposed to know that anything has gone wrong, if the underlying OS doesn't tell? Heh, magic I guess. If I mount the drive as a CIFS drive from a Linux box, I can delete the files just fine, so for now that gives me a workaround (I'll move my deletion process to a Linux box). --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 10:50 PM -0400 Larry Hall (Cygwin) reply-to-list-only...@cygwin.com wrote: Yeah, there's been plenty of these network appliances that really foul up their implementation of permissions. MVFS and NetApp are a couple in the past that have had problems. If you have the csih package installed, run /usr/lib/csih/getVolInfo drive letter:/ and send the results here. And then wait for the groans. ;-) Ok, here is the results. ;) $ /usr/lib/csih/getVolInfo /cygdrive/z Device Type: 7 Characteristics: 10 Volume Name: 121 Serial Number : 27 Max Filenamelength : 255 Filesystemname : NTFS Flags : 4007e FILE_CASE_SENSITIVE_SEARCH : FALSE FILE_CASE_PRESERVED_NAMES : TRUE FILE_UNICODE_ON_DISK: TRUE FILE_PERSISTENT_ACLS: TRUE FILE_FILE_COMPRESSION : TRUE FILE_VOLUME_QUOTAS : TRUE FILE_SUPPORTS_SPARSE_FILES : TRUE FILE_SUPPORTS_REPARSE_POINTS: FALSE FILE_SUPPORTS_REMOTE_STORAGE: FALSE FILE_VOLUME_IS_COMPRESSED : FALSE FILE_SUPPORTS_OBJECT_IDS: FALSE FILE_SUPPORTS_ENCRYPTION: FALSE FILE_NAMED_STREAMS : TRUE FILE_READ_ONLY_VOLUME : FALSE FILE_SEQUENTIAL_WRITE_ONCE : FALSE FILE_SUPPORTS_TRANSACTIONS : FALSE --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Oddities with file deletion on CIFS drive
I have a CIFS drive I connect to as the windows user. I can write to the drive with no problem. However, when I go to delete files from the drive, Cygwin behaves very oddly. bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ rm -f * bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ ls -l total 104 -r-xr-xr-x 1 1362 2010-09-08 13:31 BUILD_EVO_template -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_ISYNC_template -r-xr-xr-x 1 636 2010-09-08 13:31 BUILD_LABEL_template -r-xr-xr-x 1 1536 2010-09-08 13:31 BUILD_WIN_UPDATE_template -r-xr-xr-x 1 1699 2010-09-08 13:31 BUILD_template -r-xr-xr-x 1 4508 2010-09-08 13:31 BUILD_template_FOSS -r-xr-xr-x 1 7118 2010-09-08 13:31 BUILD_template_FOSS_ThirdParty -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_template_ISYNC -r-xr-xr-x 1 5535 2010-09-08 13:31 BUILD_template_NETWORK -r-xr-xr-x 1 7975 2010-09-08 13:31 BUILD_template_NETWORK_ThirdParty -r-xr-xr-x 1 1463 2010-09-08 13:31 BUILD_template_TOASTER -r-xr-xr-x 1 2989 2010-09-08 13:31 BUILD_template_ZDESKTOP -r-xr-xr-x 1 1386 2010-09-08 13:31 BUILD_test Even with -f, it failed to delete these files. Files that were created by this very user. Now, if I modify the file to be +w, then -w, so it returns to its original permissions, I can suddenly delete it: bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ chmod a+w BUILD_EVO_template bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ chmod a-w BUILD_EVO_template bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ rm BUILD_EVO_template bu...@zre-win-002 /cygdrive/z/current/WINDOWS/main/20100908131458_ZDESKTOP/ZimbraBuild/templates $ ls -l total 96 -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_ISYNC_template -r-xr-xr-x 1 636 2010-09-08 13:31 BUILD_LABEL_template -r-xr-xr-x 1 1536 2010-09-08 13:31 BUILD_WIN_UPDATE_template -r-xr-xr-x 1 1699 2010-09-08 13:31 BUILD_template -r-xr-xr-x 1 4508 2010-09-08 13:31 BUILD_template_FOSS -r-xr-xr-x 1 7118 2010-09-08 13:31 BUILD_template_FOSS_ThirdParty -r-xr-xr-x 1 1453 2010-09-08 13:31 BUILD_template_ISYNC -r-xr-xr-x 1 5535 2010-09-08 13:31 BUILD_template_NETWORK -r-xr-xr-x 1 7975 2010-09-08 13:31 BUILD_template_NETWORK_ThirdParty -r-xr-xr-x 1 1463 2010-09-08 13:31 BUILD_template_TOASTER -r-xr-xr-x 1 2989 2010-09-08 13:31 BUILD_template_ZDESKTOP -r-xr-xr-x 1 1386 2010-09-08 13:31 BUILD_test This behavior is quite bizarre. I should be able to delete the files I created with the -f option to rm. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 3:17 PM -0700 Quanah Gibson-Mount wrote: This behavior is quite bizarre. I should be able to delete the files I created with the -f option to rm. Also, if I mount it on a linux box via CIFS, I can delete files on this drive no issue. So it is definitely a Cygwin issue. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 6:37 PM -0400 Larry Hall (Cygwin) wrote: $ ls -l total 104 -r-xr-xr-x 1 1362 2010-09-08 13:31 BUILD_EVO_template snip I think you need to look at why the user that created the files in the first place isn't known to Cygwin. If you can solve that problem, you may find the rest falls into place. No clue, I assume it is some bug in cygwin, since this happens to any file created on this drive via Cygwin. bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ whoami build bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ ls -l blah ls: cannot access blah: No such file or directory bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ touch blah l bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ ls -l blah -rw-r--r-- 1 0 2010-09-08 15:45 blah bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ rm blah bu...@zre-win-002 /cygdrive/z/current/WINDOWS $ --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 6:51 PM -0400 Larry Hall (Cygwin) wrote: OK, take a look at http://cygwin.com/cygwin-ug-net/ntsec.html#ntsec-ids. You'll want to use 'mkpasswd' and 'mkgroup' to get the passwd and group files fixed up. I've read that page multiple times, and it still looks greek to me. :/ What I know is, the CIFS drive is mounted as the user build. The user I log into windows with is build. The user cygwin runs under is build. So all 3 of those match. I don't see why there's any issue at all determining who owns the files. bu...@zre-win-002 ~ $ id build uid=503(build) gid=513(None) groups=513(None) bu...@zre-win-002 ~ $ grep build /etc/passwd build:unused:503:513:U-ZRE-WIN-002\build,S-1-5-21-1229272821-2049760794-1417001333-1003:/home/build:/bin/bash bu...@zre-win-002 ~ $ grep 513 /etc/group None:S-1-5-21-1229272821-2049760794-1417001333-513:513: --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 9:55 PM -0400 Larry Hall (Cygwin) wrote: bu...@zre-win-002 ~ $ id build uid=503(build) gid=513(None) groups=513(None) bu...@zre-win-002 ~ $ grep build /etc/passwd build:unused:503:513:U-ZRE-WIN-002\build,S-1-5-21-1229272821-2049760794- 1417001333-1003:/home/build:/bin/bash This shouldn't be significant but is there a reason that you changed your uid to 503 from 1003? 503 is what we usually use for the build user on our linux boxes, I had changed it to see if it would do anything ownership wise, but it didn't. I changed it back to 1003 and rebooted, but no change in behavior. bu...@zre-win-002 ~ $ grep 513 /etc/group None:S-1-5-21-1229272821-2049760794-1417001333-513:513: Hm, what does 'ls -ln' look like? I can guess but I shouldn't do that. ;-) -r-xr-xr-x 1 4294967295 4294967295 1453 2010-09-08 13:31 BUILD_ISYNC_template (etc) That covers all the obvious things for me. I'd recommend taking a look at the SAMBA server to see if user IDs are mapped correctly. Also, I know there were some problems in the past with older Samba servers, bugs and all. Make sure you're using a current version there. It may be worthwhile to look at the permissions and owners from the Windows perspective too. If that doesn't get you anywhere, I'd recommend filing a full problem report on your follow-up to the list. See the link below for full details: I don't think the CIFS server is using Samba, actually. It is a Celerra storage array from EMC (http://www.emc.com/products/family/celerra-family.htm), and in this case, it isn't hooked up to an AD server like they usually have them. Problem reports: http://cygwin.com/problems.html Ok, thanks. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Oddities with file deletion on CIFS drive
--On Wednesday, September 08, 2010 7:12 PM -0700 Quanah Gibson-Mount qua...@zimbra.com wrote: That covers all the obvious things for me. I'd recommend taking a look at the SAMBA server to see if user IDs are mapped correctly. Also, I know there were some problems in the past with older Samba servers, bugs and all. Make sure you're using a current version there. It may be worthwhile to look at the permissions and owners from the Windows perspective too. If that doesn't get you anywhere, I'd recommend filing a full problem report on your follow-up to the list. See the link below for full details: I don't think the CIFS server is using Samba, actually. It is a Celerra storage array from EMC (http://www.emc.com/products/family/celerra-family.htm), and in this case, it isn't hooked up to an AD server like they usually have them. Hm, looks like it does use smb, based off google. But I have no idea what version of Samba it ships with. I'll see if I can find out from out networking team. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Saturday, September 04, 2010 11:30 AM +0200 Corinna Vinschen corinna-cyg...@cygwin.com wrote: The login via ssh is another session. The Windows mounts are only stored on a per-session base. I thought the point of using passwd -R is so that when you log in to a new session, the drive gets automatically mounted... In any case, I modified my .bashrc to mount the drive if it detects an SSH session. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Tuesday, August 24, 2010 1:35 PM -0700 Quanah Gibson-Mount wrote: --On Tuesday, August 24, 2010 1:19 PM -0400 Larry Hall (Cygwin) wrote: A missing password is likely your problem, since that's at the heart of your original problem really. But I actually wasn't thinking of this option as the solution to your problem when I pointed you at the FAQ. Running a service as the user you'll log in as is really just a workaround. I was thinking more of: http://cygwin.com/cygwin-ug-net/ntsec.html#ntsec-nopasswd3 Using the passwd -R bit did not change the behavior. I still am unable to access the network share when logging in via SSH, but it works just fine when starting cygwin locally. I will note that the network share (CIFS) uses a different username/password than my local user/password on the windows box (That, unfortunately, is not under my control). However, I set up the share using net use \\IP\share /savecreds, and that does mount the drive to Windows at least every time without prompting, and Cygwin locally is able to use it. It is only an issue with the SSH connection. The CIFS mount now uses the same username/password as the windows user. I have used passwd -R to store the user's password in the registry. The drive again shows up when launching cygwin manually, and it still fails to show up when I use ssh to connect. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Friday, September 03, 2010 12:28 PM -0700 Quanah Gibson-Mount wrote: --On Tuesday, August 24, 2010 1:35 PM -0700 Quanah Gibson-Mount wrote: The CIFS mount now uses the same username/password as the windows user. I have used passwd -R to store the user's password in the registry. The drive again shows up when launching cygwin manually, and it still fails to show up when I use ssh to connect. $ net use New connections will be remembered. Status Local RemoteNetwork --- Unavailable Z:\\10.137.242.250\Zbuild3 Microsoft Windows Network The command completed successfully. $ net use Z: '\\10.137.242.250\Zbuild3' The command completed successfully. $ net use New connections will be remembered. Status Local RemoteNetwork --- OK Z:\\10.137.242.250\Zbuild3 Microsoft Windows Network The command completed successfully. So, mounting it works ok if I do it manually via ssh. But why isn't it simply showing up when it's already mounted on the system, and the credentials have been stored via passwd -R? It was also mounted using /savecred --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Monday, August 23, 2010 9:41 PM -0400 Larry Hall (Cygwin) reply-to-list-only...@cygwin.com wrote: On 8/23/2010 7:47 PM, Quanah Gibson-Mount wrote: I have a network drive set up to auto-reconnect at startup, which appears to work fine. If I start cygwin after logging in, I can see the drive in the /cygwin directory. However, I cannot see the drive in the /cygwin directory when connecting via Cygwin's OpenSSH. My guess is that ssh starts before the drive shows up. Running net use from the SSH connection says it is unavailable: snip The FAQ is your friend: http://cygwin.com/faq/faq-nochunks.html#faq.using.shares Thanks, I missed that, was reading the fstab etc pages. So I set up sshd to run as the build user, but it still doesn't work: I ran: cygrunsrv.exe -R sshd confirmed no sshd processes were running rebooted, confirmed no sshd processes were running Ran: ssh-host-config -u build to set it up to run as the build user (my primary user) Told it to use privilege separation Said I want sshd to run as a service Put ntsec tty as the values for the CYGWIN environment variable Ran cygrunsrv -S sshd ps -eaf now shows sshd running as SYSTEM proc 3376 with parent 3340, and a subprocess sshd 3444 running as build user I log in as the build user via ssh, and ps -eaf shows me to sshd process running as SYSTEM - proc 3376 with parent 3340 and a subprocess 1824 with 3376 as the parent. So, in reading the further documents, is what I need to do then is change the SYSTEM entry in /etc/passwd to use the SID of the build user instead? --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Tuesday, August 24, 2010 9:59 AM -0700 Quanah Gibson-Mount qua...@zimbra.com wrote: I log in as the build user via ssh, and ps -eaf shows me to sshd process running as SYSTEM - proc 3376 with parent 3340 and a subprocess 1824 with 3376 as the parent. Specifically, when logging in via remote ssh, both processes were owned by SYSTEM. So the bit to run as the build user apparently doesn't actually work. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Tuesday, August 24, 2010 1:19 PM -0400 Larry Hall (Cygwin) reply-to-list-only...@cygwin.com wrote: A missing password is likely your problem, since that's at the heart of your original problem really. But I actually wasn't thinking of this option as the solution to your problem when I pointed you at the FAQ. Running a service as the user you'll log in as is really just a workaround. I was thinking more of: http://cygwin.com/cygwin-ug-net/ntsec.html#ntsec-nopasswd3 Using the passwd -R bit did not change the behavior. I still am unable to access the network share when logging in via SSH, but it works just fine when starting cygwin locally. I will note that the network share (CIFS) uses a different username/password than my local user/password on the windows box (That, unfortunately, is not under my control). However, I set up the share using net use \\IP\share /savecreds, and that does mount the drive to Windows at least every time without prompting, and Cygwin locally is able to use it. It is only an issue with the SSH connection. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Network drives ssh access
--On Tuesday, August 24, 2010 4:49 PM -0400 Larry Hall (Cygwin) wrote: However, I set up the share using net use \\IP\share /savecreds, and that does mount the drive to Windows at least every time without prompting, and Cygwin locally is able to use it. It is only an issue with the SSH connection. Well, that's a completely different thing then. You need Windows to authenticate that share as the other user then. I think you're stuck with net use option of the FAQ. You can either invoke it each time you log in via ssh, put it in your favorite rc file, or have a separate script to do it (which you'd invoke each time). I've yet to get net use to work for me via cygwin. net use Z: \\W.X.Y.Z\Share password tries to prompt me with a question: Z: has a remembered connection to \\W.X.Y.Z\Share. Do you want to overwrite the remembered connection? (Y/N) [Y]: No valid response was provided I.e., it never gives me the opportunity to answer the question. If I try changing the drive letter, I get error 67, the network name cannot be found. Note that net use by itself continues to show the Unavailable status for the Z: drive with that share. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Network drives ssh access
I have a network drive set up to auto-reconnect at startup, which appears to work fine. If I start cygwin after logging in, I can see the drive in the /cygwin directory. However, I cannot see the drive in the /cygwin directory when connecting via Cygwin's OpenSSH. My guess is that ssh starts before the drive shows up. Running net use from the SSH connection says it is unavailable: bu...@zre-win-002 / $ net use New connections will be remembered. Status Local RemoteNetwork --- Unavailable Z:\\X.X.X.X\Zbuild3 Microsoft Windows Network The command completed successfully. From the system itself: bu...@zre-win-002 / $ net use New connections will be remembered. Status Local RemoteNetwork --- OK Z:\\X.X.X.X\Zbuild3 Microsoft Windows Network The command completed successfully. Trying to run net use Z: doesn't help any. Any ideas on what I can do to make this drive show up so I can access it when accessing the windows box remotely? Thanks, Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc Zimbra :: the leader in open source messaging and collaboration -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple