Michael Hudson wrote:
> Martin Pool wrote:
>> I believe mwhudson has started working on the bug asking for a memory
>> limit on bzr processes. When that lands we should start seeing
>> failures if things do get into a bloated state, and hopefully the
>> failure will include some meaningful traceback.
>
> I doubt the approach I've taken -- setrlimit -- will result in
> tracebacks, I think it will be termination with rather more prejudice
> than that. But maybe, we'll see.
>
> Cheers,
> mwh
>
Could you do something silly like spawn a thread at startup that polls
vmstat and issues SIGINT to the running process if it exceeds a soft
threshold, and then have a hard threshold above that?
John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Michael Hudson wrote:
> Martin Pool wrote:
>> I believe mwhudson has started working on the bug asking for a memory
>> limit on bzr processes. When that lands we should start seeing
>> failures if things do get into a bloated state, and hopefully the
>> failure will include some meaningful traceback.
>
> I doubt the approach I've taken -- setrlimit -- will result in
> tracebacks, I think it will be termination with rather more prejudice
> than that. But maybe, we'll see.
>
> Cheers,
> mwh
>
Could you do something silly like spawn a thread at startup that polls
vmstat and issues SIGINT to the running process if it exceeds a soft
threshold, and then have a hard threshold above that?
John
=:->
-----BEGIN PGP SIGNATURE----- enigmail. mozdev. org/
kWzAACgkQJdeBCY SNAANbfACdHIW11 WqDOUXTxvnd1a1L cGHf f07/Hov36HMauD7 4f
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://
iEYEARECAAYFAku
LX0AoIlup91SOAO
=B52z
-----END PGP SIGNATURE-----