Paramiko is at version 1.12.0 Also, for reference the version of Stash that we are running is 2.6.4.
-Erika On Wednesday, January 8, 2014 6:07:36 PM UTC-5, Christian Hammond wrote: > > Yeah, that was a typo. > > Okay.. Well, this sounds like a compatibility issue between Paramiko and > your SSH server. > > Can you tell me what version of Paramiko is installed? You should be able > to do a 'locate paramiko' and find the egg directory. > > Christian > > -- > Christian Hammond - chi...@chipx86.com <javascript:> > Review Board - http://www.reviewboard.org > Beanbag, Inc. - http://www.beanbaginc.com > > > On Wed, Jan 8, 2014 at 2:04 PM, Erika Redmark > <maniacmon...@gmail.com<javascript:> > > wrote: > >> Thank you for the steps. I revoked the path to a local repo and followed >> the steps you gave me. Here are the results. >> >> 1) The user is www-data, and they have access to >> /var/www/reviewboard/data/.ssh, but there is only an id_rsa, not an id_dsa >> 2) I used the id_rsa key I found there as my own for my user to test. I >> was able to clone fine. >> 3) Is GIT_SSH=rbsh a mistype for GIT_SSH=rbssh? Assuming it is, I >> executed that command, and it *almost *worked: >> >> Cloning into 'plugin-suite'... >> remote: Counting objects: 20628, done. >> remote: Compressing objects: 100% (12134/12134), done. >> remote: Total 20628 (delta 9444), reused 15477 (delta 5538) >> fatal: The remote end hung up unexpectedly >> fatal: early EOF >> fatal: index-pack failed >> >> I did a tail on the rbssh log, trimming out a long line of ! ! >> process_channel stuff and got: >> 01-08 16:57 root DEBUG !! process_channel >> >> 01-08 16:57 root DEBUG !! process_channel >> >> 01-08 16:57 paramiko.transport DEBUG [chan 1] EOF received (1) >> 01-08 16:57 root DEBUG !! process_channel >> >> 01-08 16:57 root DEBUG !!! exit_status_ready >> >> 01-08 16:57 root DEBUG !!! Done >> 01-08 16:57 paramiko.transport DEBUG [chan 1] EOF sent (1) >> 01-08 16:57 paramiko.transport DEBUG Dropping user packet because >> connection is dead. >> 01-08 16:57 paramiko.transport DEBUG Dropping user packet because >> connection is dead. >> >> Would you also need to see the beginning of the file? >> >> - Erika >> >> On Wednesday, January 8, 2014 4:24:25 PM UTC-5, Christian Hammond wrote: >> >>> Hmm, I thought we had turned off validation when the URL mask was >>> supplied. I guess I was wrong. >>> >>> So I'd have expected the log to have said something one way or another >>> about keys, whether it had one or didn't, and whether or not it succeeded. >>> An "EOF in transport thread" usually means something else went wrong. We'll >>> need to figure out what that is. >>> >>> In this particular case, we're not using rbssh for the communication, >>> but rather Paramiko (the library rbssh wraps), which is being used from >>> Review Board itself. This fortunately means we don't have to mess around >>> with getting log output from rbssh, but also means what we're seeing is >>> about as good as it gets from the logging. >>> >>> Can you try a few things? >>> >>> 1) Make sure that /path/to/site/data/.ssh/id_dsa is accessible by the >>> web server. >>> 2) Temporarily use this as your own SSH key and try to do a clone with >>> that repository path, to make sure that a clone with actual git + SSH works >>> as expected. >>> 3) Then, do the same thing with: GIT_SSH=rbsh DEBUG_RBSSH=1 git clone >>> ...... >>> >>> #2 should work so long as your configuration is correct on the Git >>> server. If it fails, something is wrong there. >>> >>> If #2 works and #3 fails, then look in the current directory for ab >>> rbssh*.log file. See if it has anything more useful to say. >>> >>> Christian >>> >>> -- >>> Christian Hammond - chi...@chipx86.com >>> Review Board - http://www.reviewboard.org >>> Beanbag, Inc. - http://www.beanbaginc.com >>> >>> >>> On Wed, Jan 8, 2014 at 1:10 PM, Erika Redmark <maniacmon...@gmail.com>wrote: >>> >>>> I read about the issue involving git and I had previously used a local >>>> clone, but ended up with errors where it would not find diff changes for >>>> new branches. I do have a mask I can use that works with Stash: >>>> >>>> https://[redacted]/projects/key/repos/plugin-suite/browse/ >>>> <filename>?at=<revision>&raw >>>> >>>> But the error comes up whether I use a mask or not when I supply the >>>> actually remote location for the 'Path' field. I have two options: SSH and >>>> HTTPS. The error I get above is when I use the SSH clone uri. I simply >>>> just >>>> get a "Repository Not Found" if I use the HTTPS clone uri (along with the >>>> proper password for my user). >>>> >>>> It doesn't matter if I provide the mask or not. The error is occurring >>>> based on what I supply to Path. If I leave it blank and just supply the >>>> mask, it tells me Path is a required field. >>>> >>>> On Wednesday, January 8, 2014 3:40:01 PM UTC-5, Christian Hammond wrote: >>>> >>>>> Hi Erika, >>>>> >>>>> So the first big thing to know about Git is that you *must* either use >>>>> a local clone, or provide the raw file mask. Ideally the latter. Git has >>>>> no >>>>> protocol support for fetching the content of an individual file at a >>>>> given >>>>> revision, which means you can't just provide a remote path to Review >>>>> Board >>>>> and have it work. This is truly annoying, but there's little we can do >>>>> about it. >>>>> >>>>> So, the thing to focus on is figuring out if there's a URL you can use >>>>> to fetch the contents of a file at a path + revision that Review Board >>>>> can >>>>> access through either anonymous access or using Basic HTTP auth. If there >>>>> is, plug that in and things should work, but if not, you'll need another >>>>> service that can provide that for you. >>>>> >>>>> Basically, rbssh shouldn't really factor into this at this point. >>>>> >>>>> Christian >>>>> >>>>> -- >>>>> Christian Hammond - chi...@chipx86.com >>>>> >>>>> Review Board - http://www.reviewboard.org >>>>> Beanbag, Inc. - http://www.beanbaginc.com >>>>> >>>>> >>>>> On Wed, Jan 8, 2014 at 7:09 AM, Erika Redmark >>>>> <maniacmon...@gmail.com>wrote: >>>>> >>>>>> This question is almost identical to https://groups.google.com/f >>>>>> orum/#!topic/reviewboard/UmXYqdsKTXY, however that thread ended >>>>>> without resolution. >>>>>> >>>>>> We have ReviewBoard 1.7.20 set up on Ubuntu 12.04. I am trying to set >>>>>> up a repository on ReviewBoard against a git repository on stash, but I >>>>>> run >>>>>> into the "[Errno -5] No address associated with hostname" error. This >>>>>> happens when I have a valid ssh clone link in my Path field. For this I >>>>>> am >>>>>> leaving all the other fields empty. My log is almost identical to the >>>>>> one >>>>>> posted on the other topic. I have redacted the actual links to our >>>>>> server. >>>>>> >>>>>> GitTool: Attempting ssh connection with host: [redacted]:7999, >>>>>> username: git >>>>>> 2014-01-08 14:22:17,188 - DEBUG - - starting thread (client mode): >>>>>> 0xb9b412ecL >>>>>> 2014-01-08 14:22:17,208 - INFO - - Connected (version 2.0, client >>>>>> SSHD-CORE-0.7.0-ATLASSIAN-M2) >>>>>> 2014-01-08 14:22:17,236 - DEBUG - - kex algos:['diffie-hellman- >>>>>> group14-sha1', 'diffie-hellman-group1-sha1'] server key:['ssh-rsa'] >>>>>> client encrypt:['aes128-cbc', '3des-cbc', 'blowfish-cbc'] server >>>>>> encrypt:['aes128-cbc', '3des-cbc', 'blowfish-cbc'] client >>>>>> mac:['hmac-md5', >>>>>> 'hmac-sha1', 'hmac-md5-96', 'hmac-sha1-96'] server mac:['hmac-md5', >>>>>> 'hmac-sha1', 'hmac-md5-96', 'hmac-sha1-96'] client compress:['none'] >>>>>> server >>>>>> compress:['none'] client lang:[''] server lang:[''] kex follows?False >>>>>> 2014-01-08 14:22:17,236 - DEBUG - - Ciphers agreed: >>>>>> local=aes128-cbc, remote=aes128-cbc >>>>>> 2014-01-08 14:22:17,236 - DEBUG - - using kex >>>>>> diffie-hellman-group1-sha1; server key type ssh-rsa; cipher: local >>>>>> aes128-cbc, remote aes128-cbc; mac: local hmac-sha1, remote hmac-sha1; >>>>>> compression: local none, remote none >>>>>> 2014-01-08 14:22:17,371 - DEBUG - - Switch to new keys ... >>>>>> 2014-01-08 14:22:17,472 - DEBUG - - EOF in transport thread >>>>>> >>>>>> As in the other topic, initially this looks like a key problem >>>>>> because it never once says anything about key authenticated, or key >>>>>> denied; >>>>>> it just ends. However, the line: >>>>>> >>>>>> Connected (version 2.0, client SSHD-CORE-0.7.0-ATLASSIAN-M2) >>>>>> >>>>>> It sees this is stash, so I gather that it is properly getting to the >>>>>> right address (otherwise it wouldn't know). So having verified that is >>>>>> actually getting to the host in question, things go back to the keys. >>>>>> >>>>>> On Stash, I made a user, let's call him ReviewBoardUser. On >>>>>> ReviewBoard, I generated a key using the SSH tab on the left. I took >>>>>> this >>>>>> key and gave it to the ReviewBoardUser. What I expect would happen is >>>>>> that >>>>>> ReviewBoard would authenticate with Stash with the key, Stash would map >>>>>> it >>>>>> to this special ReviewBoardUser that has read access to all repositories >>>>>> that ReviewBoard needs, and then everything would be fine. (On that >>>>>> note, I >>>>>> would be using the form >>>>>> "https://server/path/to/repo/<filename>?at=<revision>&raw" >>>>>> as the mask for Stash, so I don't believe there would be a need for cgit >>>>>> or >>>>>> anything. I won't know until I get this working and can actually try it). >>>>>> >>>>>> The previous topic user was also using Stash from the looks of the >>>>>> logs. >>>>>> >>>>>> One last point regarding the previous topic; I am not sure how to use >>>>>> rbssh properly, so I couldn't get any extra info from using that tool. I >>>>>> would appreciate a nice example syntax for using it, as I can't tell if >>>>>> I >>>>>> am making an error, or if it is itself reporting a valid error with the >>>>>> configuration. >>>>>> >>>>>> Let me know if there is any more information that is needed, I will >>>>>> be happy to provide it. >>>>>> >>>>>> - Erika >>>>>> >>>>>> -- >>>>>> Get the Review Board Power Pack at http://www.reviewboard.org/pow >>>>>> erpack/ >>>>>> --- >>>>>> Sign up for Review Board hosting at RBCommons: https://rbcommons.com/ >>>>>> --- >>>>>> Happy user? Let us know at http://www.reviewboard.org/users/ >>>>>> --- >>>>>> You received this message because you are subscribed to the Google >>>>>> Groups "reviewboard" group. >>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>> send an email to reviewboard...@googlegroups.com. >>>>>> >>>>>> For more options, visit https://groups.google.com/groups/opt_out. >>>>>> >>>>> >>>>> -- >>>> Get the Review Board Power Pack at http://www.reviewboard.org/ >>>> powerpack/ >>>> --- >>>> Sign up for Review Board hosting at RBCommons: https://rbcommons.com/ >>>> --- >>>> Happy user? Let us know at http://www.reviewboard.org/users/ >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "reviewboard" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to reviewboard...@googlegroups.com. >>>> For more options, visit https://groups.google.com/groups/opt_out. >>>> >>> >>> -- >> Get the Review Board Power Pack at http://www.reviewboard.org/powerpack/ >> --- >> Sign up for Review Board hosting at RBCommons: https://rbcommons.com/ >> --- >> Happy user? Let us know at http://www.reviewboard.org/users/ >> --- >> You received this message because you are subscribed to the Google Groups >> "reviewboard" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to reviewboard...@googlegroups.com <javascript:>. >> For more options, visit https://groups.google.com/groups/opt_out. >> > > -- Get the Review Board Power Pack at http://www.reviewboard.org/powerpack/ --- Sign up for Review Board hosting at RBCommons: https://rbcommons.com/ --- Happy user? Let us know at http://www.reviewboard.org/users/ --- You received this message because you are subscribed to the Google Groups "reviewboard" group. To unsubscribe from this group and stop receiving emails from it, send an email to reviewboard+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.