Hello,
I am using the latest Pharo based Pier one click image on debian server, and I have set up Image persistency. Problem is that backup images seem to be cumulating and eating my disk space quite rapidly. As far as I have understood, there should be a max of 3 backed images, while I have tens, and even hundreds of them. I suspected that maybe it is the file rights issue, but everything seems to be ok, (also if a process can create file, it usually can also delete it). Thanks in advance! rush http://www.cloud208.com/ _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
The maximum number of images can be set in the configuration dialog in
"/systemmanagement/persistency", the default is 3. To debug the deletion of old images you can debug the following expression: PRKernel instances anyOne persistency cleanupBackupImages Cheers, Lukas On Tue, Apr 14, 2009 at 10:11 AM, Davorin Rusevljan <[hidden email]> wrote: > Hello, > > I am using the latest Pharo based Pier one click image on debian > server, and I have set up Image persistency. > > Problem is that backup images seem to be cumulating and eating my disk > space quite rapidly. As far as I have understood, there should be a > max of 3 backed images, while I have tens, and even hundreds of them. > > I suspected that maybe it is the file rights issue, but everything > seems to be ok, (also if a process can create file, it usually can > also delete it). > > Thanks in advance! > > rush > http://www.cloud208.com/ > _______________________________________________ > Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki > -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
On 4/14/09, Lukas Renggli <[hidden email]> wrote:
> The maximum number of images can be set in the configuration dialog in > "/systemmanagement/persistency", the default is 3. yes I have also that set to 3. > To debug the deletion of old images you can debug the following expression: > > PRKernel instances anyOne persistency cleanupBackupImages Interestingly, this also seems to be running fine when started alone, it had cleaned all old files and kept only last 3. So something before that (but after the image save) must be breaking persistency cycle. Thanks! rush http://www.cloud208.com/ _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
>> To debug the deletion of old images you can debug the following expression:
>> >> PRKernel instances anyOne persistency cleanupBackupImages > > Interestingly, this also seems to be running fine when started alone, > it had cleaned all old files and kept only last 3. So something before > that (but after the image save) must be breaking persistency cycle. I can't reproduce on my machine and on various servers. What happens if you repeatedly evaluate the following expression? PRKernel instances anyOne persistency snapshot This is the code that is executed whenever the image is saved. It should create a backup copy of the current image, save the image under the same name, and delete older images. Does that work for you if you evaluate the expression repeatedly? What version of Squeak and Pier are you working with? Lukas -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
> This is the code that is executed whenever the image is saved. It
> should create a backup copy of the current image, save the image under > the same name, and delete older images. Does that work for you if you > evaluate the expression repeatedly? PRKernel instances anyOne persistency snapshot raises error every time I run it, her is TOS: ------- Stack Trace thisContext HttpService(Object)>>error: self seaside [stopped] port: 8080 aString 'a service is already running on port 8080' ------------- before that it saves image ok, and pier image keeps responding to requests afterwards. I am using latest (1.1.1) pier one-click image based on pharo, downloaded from www.piercms.org Probable is my pier app config, that is triggering the problem, i have: -------- Server Resource Base Url (unspecified) Server Hostname clear Server Path clear Server Port revert to: 80 overriden from WAGlobalConfiguration Server Protocol #http override inherited from WAGlobalConfiguration ------- notice the port:80, this is the port that nginx responds to, and forwards requests to pier image at port:8080 rush http://www.cloud208.com/ _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
> before that it saves image ok, and pier image keeps responding to
> requests afterwards. Aha, now I understand. You have multiple web-server adapters listening on the same port. After the image is saved they restart and throw an exception. Stop them all using: WAKom withAllSubclassesDo: #stop And then start your favorite adapter using: WAKom startOn: 8080 That should solve the problem. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
Ok, here is what I get. (I am doing this through Halo inspector web interface) WAKom withAllSubclassesDo: #stop . WAKomEncoded startOn: 8080 . -- runs fine.
TcpService services (click on inspect buton) -- nice only one Seside running service returned on port 8080. PRKernel instances anyOne persistency snapshot -- bangs with exception stack described few messages above.
TcpService services -- not nice, now has 2 Seaside services on port 8080 one of them stoped. thanks! rush _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
> -- bangs with exception stack described few messages above.
> TcpService services Did you try stopping "TcpService services" as well? Lukas -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
On Wed, Apr 15, 2009 at 8:27 AM, Lukas Renggli <[hidden email]> wrote:
ok, now I have tried: WAKom withAllSubclassesDo: #stop . TcpService services do: [:s | s stop]. WAKomEncoded startOn: 8080 .
-- ok, only one service left running on port 8080 PRKernel instances anyOne persistency snapshot -- again same exception, also has 2 services on port 8080, one stoped other running.
So I think snapshot is somehow generating more than one service for port 8080. Also here is the Cloud208.sh script used to start pier if that migh give you some clue:
----------------- #!/bin/sh dir="/ebs/cloud208/Pier" EXE="$dir/Contents/Linux686" RES="$dir/Contents/Resources" case "$1" in start) nohup su -c "$EXE/squeak -plugins $EXE -vm-display-null -vm-soun d-null -mmap 100m $RES/pier.image '' port 8080" squeak &
echo $! > $dir/pid ;; stop) kill `cat $dir/pid` ;; *) exit 3
;; esac ----------------- thanks! rush _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
> d-null -mmap 100m $RES/pier.image '' port 8080" squeak &
> echo $! > $dir/pid > ;; Looks like you should remove the "port 8080" thing. Lukas -- Lukas Renggli http://www.lukas-renggli.ch _______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
On Wed, Apr 15, 2009 at 12:00 PM, Lukas Renggli <[hidden email]> wrote:
yes, that was that. many thanks, and sorry for the trouble! rush
_______________________________________________ Magritte, Pier and Related Tools ... https://www.iam.unibe.ch/mailman/listinfo/smallwiki |
Free forum by Nabble | Edit this page |