jotobjects
|
Posted: May 06, 2010 14:41 by jotobjects
|
|
Having a problem where "hg pull" and "hg clone" are suddenly not working. Was working until yesterday. hg pull (https) hangs on "adding file changes". Same thing happens exactly with hg clone. Here is what I see on the screen - pulling from https://hg.kenai.com/hg/jfxphoneapps~javafxphoneapps searching for changes adding changesets adding manifests adding file changes I have left it there for hours - nothing happens - no CPU usage or network traffic after the first few seconds. Any ideas on what might the problem? hg verify has no errors on local repo. |
HG mercurial hangs
8 posts
Replies: 7 - Last Post: May 14, 2010 19:40
by: jotobjects
by: jotobjects
showing 1 - 8 of 8
Edwin Goei
|
Posted: May 06, 2010 17:28 by Edwin Goei
|
| As a workaround, have you tried ssh access? Using ssh is what I would recommend switching to anyway at least for read-write access. For more info on repo access alternatives, see the table on http://blogs.sun.com/edwingo/entry/kenai_com_source_repository_access. I will look at the https access problem as that should work. |
|
jotobjects
|
Posted: May 06, 2010 19:31 by jotobjects
|
|
Thanks - I'm outside the firewall and cannot get ssh access. We have some people using https that can use the project and some who can't - and it changes at different times for the same people - suggesting an intermittent problem. The mercurial developer on the HG users mail list suggests it is a network or server issue of some kind. This started after the move last weekend. This is a huge problem for us. A large project is coming to screeching halt as a result. Thanks for your help. |
|
jotobjects
|
Posted: May 06, 2010 19:45 by jotobjects
|
| I may be wrong about not being able to use ssh from outside firewall - will ask Lea via email |
|
jotobjects
|
Posted: May 06, 2010 22:45 by jotobjects
|
|
I did get ssh access set up - but now got this error. It seems to abort at a different place every time. hg clone ssh://myname@hg.kenai.com/jfxphoneapps~javafxphoneapps Enter passphrase for key '/cygdrive/c/Users/jot/.ssh/id_rsa': destination directory: jfxphoneapps~javafxphoneapps requesting all changes adding changesets adding manifests adding file changes transaction abort! rollback completed abort: premature EOF reading chunk (got 112426 bytes, expected 211237) remote: Read from remote host hg.kenai.com: Software caused connection abort |
|
jotobjects
|
Posted: May 10, 2010 16:44 by jotobjects
|
|
We are still stuck - Here are some messages from Matt MacKall the lead developer for mercurial. NUTSHELL - The fact that we have seen this on Windows and Linux clients make it almost certain this is a server problem. We switched from https to ssh and that did not help. Email #1 mercurial@selenic.com On 5/7/2010 1:40 PM, Matt Mackall wrote: > On Fri, 2010-05-07 at 13:25 -0700, Paul Copeland wrote: > >> Following up on prior topic "hg pull stalls" - >> >> I switched to ssh from https and was hopeful that would take care of it. No luck - Now I it quits with an error with ssh - so that's better than stalling. When I use --debug it again stops on the kind of file where it stalled with https. hg clone also fails. >> >> This is a large server with many projects. They just moved to a new datacenter. They have lots of users and they claim that only me and a few others on my project are having this problem. >> >> Does the error from hg clone below give any hints about the problem? This was all working until a few days ago. Some of the users on my project are not having this problem and some are. >> > Is there any pattern to the operating systems involved? Are any Linux > clients failing? > > Almost all problems with push/pull are on the server side, and there's > no sign that this is an exception. That it's sporadic and affects only > some clients reinforces that. It's very likely hitting a resource limit > or timeout or similar. How long does it take to fail? > > If the server folks still aren't buying it, feel free to direct them to > me. > > In the mean time, you might be able to make some forward progress using > partial pulls (see pull -r) > Email #2 mercurial@selenic.com On 5/10/2010 12:36 AM, Matt Mackall wrote: > On Mon, 2010-05-10 at 12:41 +1000, Cameron Simpson wrote: > >> On 07May2010 15:35, Paul Copeland <tech@jotobjects.com> wrote: >> | We have had 5 clients failing some intermittently and some consistently. >> [ ... both linux and windows ... ] >> | The server people are looking into it... [...] >> >> Speaking as a network/firewall/OS guy rather than a mercurial guy, >> is it possible that a stateful firewall has decided the connection is >> idle? Matt, are there periods where the hg pull might send no data in >> either direction for an extended period? >> > > On typical networks links, Mercurial is bandwidth limited. For extremely > large files (ie much bigger than we seem to be talking about here), > there may be noticeable delays. But a proxy/firewall could definitely be > to blame if it turns out to be a common factor among failing users. > > The only two common factors yet identified are a) server and b) repo. > Paul, have you tried the incremental pull thing yet? It might help > identify a consistently failing cset, or it might get you past the > problem. Email #3 mercurial@selenic.com On 5/6/2010 11:20 AM, Matt Mackall wrote: > On Thu, 2010-05-06 at 09:54 -0700, Paul Copeland wrote: > >> Thanks! - progressbar didn't help - stalled the same. >> >> With --debug these are the last few lines of output - >> >> adding JFXPhoneApps/src/common/com/oracle/satin/apps/news/SourceBox.fx revision >> >> files: 30 chunks >> files: 31 chunks >> files: 32 chunks >> adding JFXPhoneApps/src/common/com/oracle/satin/apps/news/SourceEditBox.fx revi >> ions >> files: 33 chunks >> files: 34 chunks >> adding JFXPhoneApps/src/common/com/oracle/satin/apps/news/images/arrow.png revi >> ions >> files: 35 chunks >> adding JFXPhoneApps/src/common/com/oracle/satin/apps/news/images/background.png >> revisions >> >> It stays there forever apparently. Does this mean it was working on background.png and hung there? background.png is 117 KB. I don't know if that is considered "large". We have some larger files. >> > > No, that's not large. Large would be, say 100MB or more. > > Does it stop there consistently? It looks like the answer is no as your > other email says 92 rather than 35. That strongly suggests a server > issue, like a timeout. How long does this take? > > Alternately, you may have a proxy or firewall issue - you should try > doing your pull on a machine on a different network. > > You'll probably need to bring the server people into the loop to check > logs and so on on their end. |
|
jotobjects
|
Posted: May 14, 2010 19:40 by jotobjects
|
|
Just a wrap on this problem. It was never solved. As noted by Mercurial lead developer in above email excerpts, all signs point to a server or network problem after the recent move of the Kenai servers. The persistent problems were with Windows 7 and Vista clients (intermittent problems with a Linux client also). Other users with Mac clients did not experience problems. It appears the Windows clients are less able to compensate for the new server limitations (whatever they may be). The workaround was to install VMware Player with Ubuntu Linux on the Vista machine, as follows - 1. created VMware ubuntu instance on my Vista 2. shared a directory from Vista for sharing with VM Linux 3. did hg clone on Linux to shared directory 4. now have repo on Vista and can work normally and do hg work on Linux |
showing 1 - 8 of 8
Replies: 7 - Last Post: May 14, 2010 19:40
by: jotobjects
by: jotobjects
