Forum OpenACS Development: Re: CVS monitoring for OpenACS: fisheye.openacs.org

Collapse
Posted by Mark Aufflick on
Is it just me, or is that server no longer up?
Collapse
Posted by Stefan Sobernig on
Mark,

I just noticed that our virtual machine hosting fisheye became victim of a renegade fsck scrambling. currently, the
machine is in an unbootable state.

we hope to recover the fisheye service in the days to come.

as for your questions:

- fisheye is not that demanding in terms of disk space. it depends on the mode of operation and this, in turn, depends on your target SCM repository. In case of CVS, a local replica of the repository is necessary (or simply fisheye needs to be colocated with cvs). There is no pserver et al. support. SVN is supported in remote mode. the official recommendation is that your fisheye hosting system offers 3 times the size of your repository (in the cvs case).

- initial import is time-intensive. The OpenACS cvs history took 18 hours to be imported by fisheye (according to victor). it might be much less for smaller-scale repos ...

//stefan

Collapse
Posted by Stefan Sobernig on
fisheye.openacs.org is up again!

//stefan

Collapse
Posted by Richard Hamilton on
I think it may be down now as of 1800z on 29th November 2012.
Collapse
Posted by Gustaf Neumann on
In the meantime, you might try: https://github.com/openacs
Collapse
Posted by Stefan Sobernig on
I fired up fisheye again. It turned out that the JVM (6.0_24-b24) crashed on Nov 29, around 10pm, because of a VM-internal assertion being violated:

# Internal Error (safepoint.cpp:300), pid=25655, tid=140048142944000
# guarantee(PageArmed == 0) failed: invariant

I won't invest any time to find the root cause. For now, I updated the JVM to a more recent patch version. Let's see whether this works. If not, I will attempt a more progressive upgrade of both fisheye and jvm.

//stefan

Collapse
Posted by Stefan Sobernig on
Fisheye went down again. I will need to take a closer look.

//s

Collapse
Posted by Stefan Sobernig on
fisheye is up and running again (has been for the last 20 hours).

//stefan