Differences between revisions 1 and 2
Revision 1 as of 2005-06-30 21:53:30
Size: 869
Editor: KenSchalk
Comment: New page for "don't use RH9/FC1 unless you want your repository to deadlock"
Revision 2 as of 2005-11-04 22:41:19
Size: 1440
Editor: KenSchalk
Comment: Note about the neccessity of portmap
Deletions are marked like this. Additions are marked like this.
Line 5: Line 5:
 1. [wiki:Self:VestaFAQ/Troubleshooting#deadlock Sometimes my repository or build just seems to stop]  1. [wiki:Self:VestaFAQ/TroubleShooting#deadlock Sometimes my repository or build just seems to stop]
 1. [wiki:Self:VestaFAQ/TroubleShooting#portmap Mounting the repository takes a long time]
Line 10: Line 11:
The Linux implementation of pthreads has gone through a lot of revisions. Red Hat 9 was the first Red Hat release to use the [http://en.wikipedia.org/wiki/NPTL NPTL (Native POSIX Thread Library) rewrite]. The first few OS releases which included it had some subtle bugs which are most often observed in fairly complicated multi-threaded applications such as Vesta. By Fedora Core 2 these problems seem to have been resolved.  The Linux implementation of pthreads has gone through a lot of revisions. Red Hat 9 was the first Red Hat release to use the [http://en.wikipedia.org/wiki/NPTL NPTL (Native POSIX Thread Library) rewrite]. The first few OS releases which included it had some subtle bugs which are most often observed in fairly complicated multi-threaded applications such as Vesta. By Fedora Core 2 these problems seem to have been resolved.

 Mounting the repository takes a long time:: [[Anchor(portmap)]] The Vesta repository is an NFS server. Unfortunately, the kernel will '''always''' contact the portmap(8) daemon when mounting an NFS volume, even when the port number is explicitly specified (as it is with the Vesta repository). Make sure that you have portmap installed and running on your repository server host. You can usually check this with:

{{{
rpcinfo -p `vgetconfig Repository NFS_host`
}}}

Questions in this catagory:

  1. [wiki:VestaFAQ/TroubleShooting Sometimes my repository or build just seems to stop]

  2. [wiki:VestaFAQ/TroubleShooting Mounting the repository takes a long time]


Sometimes my repository or build just seems to stop

Anchor(deadlock)Red Hat 9 and Fedora Core 1 are known to have buggy implementations of the pthreads multi-threading API. If you're running one of these, please consider upgrading to Fedora Core 2 or higher.

The Linux implementation of pthreads has gone through a lot of revisions. Red Hat 9 was the first Red Hat release to use the [http://en.wikipedia.org/wiki/NPTL NPTL (Native POSIX Thread Library) rewrite]. The first few OS releases which included it had some subtle bugs which are most often observed in fairly complicated multi-threaded applications such as Vesta. By Fedora Core 2 these problems seem to have been resolved.

Mounting the repository takes a long time

Anchor(portmap) The Vesta repository is an NFS server. Unfortunately, the kernel will always contact the portmap(8) daemon when mounting an NFS volume, even when the port number is explicitly specified (as it is with the Vesta repository). Make sure that you have portmap installed and running on your repository server host. You can usually check this with:

rpcinfo -p `vgetconfig Repository NFS_host`