Re: Zookeeper on 60+Gb mem

2010-10-05 Thread Dave Wright
cking/coordination on top. -Dave Wright On Tue, Oct 5, 2010 at 5:27 PM, Patrick Hunt wrote: > Tuning GC is going to be critical, otw all the sessions will timeout (and > potentially expire) during GC pauses. > > Patrick > > On Tue, Oct 5, 2010 at 1:18 PM, Maarten Koopmans wrote

Re: looking for info on large scale zookeeper deployments

2010-09-08 Thread Dave Wright
ient. When they initially connect, the use the cached list of servers from the last time they connected and we have a discovery mechanism for clients who have never connected. If there's interest in the modified C client we're happy to submit a patch. A similar modification to the Ja

Re: zk connect string

2010-09-03 Thread Dave Wright
No, and yes. If you don't provide all the servers it does not discovered them and won't failover it's connection If the one it is connected to fails. On Sep 3, 2010 7:58 PM, "Jun Rao" wrote: > Hi, > > If I have a replicated ZK cluster and a client is only given one of the ZK > server name in t

Re: chroot

2010-09-03 Thread Dave Wright
You need to first create the /root mode (without chroot) before connecting chrooted to it. -Dave On Sep 3, 2010 7:54 PM, "Jun Rao" wrote: > Hi, > > I set my ZK connection string to localhost:2181/root. After that, I can't > seem to be able to create path any more. Always get NodeNotExist excepti

Re: closing session on socket close vs waiting for timeout

2010-08-31 Thread Dave Wright
ssues. -Dave Wright On Tue, Aug 31, 2010 at 1:03 PM, Ted Dunning wrote: > That doesn't sound right to me. > > Is there a Zookeeper expert in the house? > > On Tue, Aug 31, 2010 at 8:58 AM, Fournier, Camille F. [Tech] < > camille.fourn...@gs.com> wrote: > > >

Re: getting created child on NodeChildrenChanged event

2010-08-31 Thread Dave Wright
27;t think it's likely to be implemented. -Dave Wright On Tue, Aug 31, 2010 at 3:49 AM, Todd Nine wrote: > Hi all, > I'm writing a distributed queue monitoring class for our leader node in > the cluster. We're queueing messages per input hardware device, this queue > is

Re: What roles do "even" nodes play in the ensamble

2010-08-25 Thread Dave Wright
ch transaction, only leader changes. -Dave Wright On Wed, Aug 25, 2010 at 6:22 PM, Todd Nine wrote: > Do I get any read performance increase (similar to an observer) since > the node will not have a voting role? > >

Re: What roles do "even" nodes play in the ensamble

2010-08-25 Thread Dave Wright
SOFTWARE ENGINEER > > todd nine| spidertracks ltd | 117a the square > po box 5203 | palmerston north 4441 | new zealand > P: +64 6 353 3395 | M: +64 210 255 8576 > E: t...@spidertracks.co.nz W: www.spidertracks.com > > > > > > On Wed, 2010-08-25 at 17:57 -0400,

Re: What roles do "even" nodes play in the ensamble

2010-08-25 Thread Dave Wright
Unfortunately ZK doesn't have any kind of dynamic configuration like that currently. You need to update all the config files and restart the ensemble. -Dave Wright

Re: Parent nodes & multi-step transactions

2010-08-23 Thread Dave Wright
to work on any single feature, that's probably the one I would pick, although I'm concerned that there would be resistance to accepting upstream. -Dave Wright On Mon, Aug 23, 2010 at 6:51 PM, Gustavo Niemeyer wrote: > Hi Mahadev, > >>  Usually the paradigm I like to s

Re: Weird ephemeral node issue

2010-08-16 Thread Dave Wright
Another possible cause for this that I ran into recently with the c client - you don't get the session expired notification until you are reconnected to the quorum and it informs you the session is lost. If you get disconnected and can't reconnect you won't get the notification. Personally I thin

Re: Using watcher for being notified of children addition/removal

2010-08-02 Thread Dave Wright
iously with 1m nodes there is a lot of overhead. I would suggest splitting your list into smaller pieces, say 1000 nodes each, and setting a watch on each of them. You mentioned using sequential nodes, which wouldn't work the same way in this case, but either way a redesign is probably in ord

Re: Regarding the process method of Watcher Interface

2010-07-19 Thread Dave Wright
There is a single thread used to fire notification callbacks, your callbacks will be called in order one at a time and you only need to worry about synchronizing state with the rest of your code. -Dave Wright On Mon, Jul 19, 2010 at 5:37 PM, Srikanth Bondalapati: wrote: > Hi, > > I hav

Re-registering watches after disconnection?

2010-05-27 Thread Dave Wright
y triggered if the watched data has changed. Although not explicitly stated, I'm assuming that in case of an actual session loss, all watches are lost and need to be re-created. -Dave Wright

Re: Dynamic adding/removing ZK servers on client

2010-05-03 Thread Dave Wright
> > > This is tricky: what happens if the server your client is connected to is > decommissioned by a view change, and you are unable to locate another server > to connect to because other view changes committed while you are > reconnecting have removed all the servers you knew about. We'd need to

Re: Dynamic adding/removing ZK servers on client

2010-05-03 Thread Dave Wright
> Should this be a znode in the privileged namespace? > I think having a znode for the current cluster members is part of the ZOOKEEPER-107 proposal, with the idea being that you could get/set the membership just by writing to that node. On the client side, you could watch that znode and update yo

Re: Dynamic adding/removing ZK servers on client

2010-05-03 Thread Dave Wright
> Hi Dave, > Just a question on how do you see it being used, meaning who would call > addserver and removeserver? It does seem useful to be able to do this. This > is definitely worth working on. You can link it as a subtask of > ZOOKEEPER-107. > In my case, it would be my client application - I

Dynamic adding/removing ZK servers on client

2010-05-03 Thread Dave Wright
ctionality will eventually be needed for whatever full dynamic server support is eventually implemented. -Dave Wright