[Glass] [GLASS] Need to restart GLASS...
Locked
[Glass] [GLASS] Need to restart GLASS...
 
|
Well, I knew it had to happen eventually. It looks like Amazon took down my instance and did some maintenance on it, so all of my GemStone processes list as Killed when I run the gslist command. I’m going to have to dig up my notes on how to get the whole system up and running again, as it has been running nonstop for something like two years, IIRC. As I’m no longer actively working on this project, I’m at a little bit of a loss as to all the things I have to do, LOL.
Anything I should be aware of when trying to bring it all back up? I have to bring up nginx, FastCGI, and all the rest. It should just restore to the last commit point in the logs, I assume. Regards, Larry _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] [GLASS] Need to restart GLASS...
|
|
The “Killed” status can be cleared by using ‘gslist -c’ to “clear” the lock files (typically in /opt/gemstone/locks). And, yes, the GemStone system will restore to the last commit point. As to the other processes, I don’t have much to contribute.
James On Sep 15, 2014, at 6:42 PM, Lawrence Kellogg via Glass <[hidden email]> wrote: > Well, I knew it had to happen eventually. It looks like Amazon took down my instance and did some maintenance on it, so all of my GemStone processes list as Killed when I run the gslist command. I’m going to have to dig up my notes on how to get the whole system up and running again, as it has been running nonstop for something like two years, IIRC. As I’m no longer actively working on this project, I’m at a little bit of a loss as to all the things I have to do, LOL. > > Anything I should be aware of when trying to bring it all back up? I have to bring up nginx, FastCGI, and all the rest. It should just restore to the last commit point in the logs, I assume. > > Regards, > > Larry > _______________________________________________ > Glass mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/glass _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
Well, it looks like I can start and stop Gemstone without any problem, as well as netldi, but I’m having trouble getting nginx to talk to my fastcgi gems. Perhaps it has something to do with my A record from Register.com, I’m not sure. I checked the IP address of my running Amazon instance, and I see that I have an elastic IP address allocated, which is in the A record on Register. Register sticks some other ip address in the first line of my A record, something I have noticed in the past. I removed that other ip address, but changes take a while to propagate.
Strangely, while the ip address seems fine in the nginx error log, the hostname is wrong, it somehow to think it is a different server than the one I’m accessing. upstream timed out (110: Connection timed out) while connecting to upstream, client: zz.zz.zz.zz, server: , request: "GET / HTTP/1.1", upstream: "fastcgi://xx.xx.xx.xx:9001", host: “www.practjour.com" (should be www.practicemusic.com) I also own www.practjour.com. I wrote a little supervise job to bring up the fast cgi gems automatically if they ever fall down. that seems to work. I changed my nginx.conf file to make sure it points to the ip address of my server, plus :9001, :9002, and :9003 for the gems. So, I’m kind of stumped at the moment. I’ll work on it some more later. Larry On Sep 15, 2014, at 9:59 PM, James Foster <[hidden email]> wrote: > The “Killed” status can be cleared by using ‘gslist -c’ to “clear” the lock files (typically in /opt/gemstone/locks). And, yes, the GemStone system will restore to the last commit point. As to the other processes, I don’t have much to contribute. > > James > > On Sep 15, 2014, at 6:42 PM, Lawrence Kellogg via Glass <[hidden email]> wrote: > >> Well, I knew it had to happen eventually. It looks like Amazon took down my instance and did some maintenance on it, so all of my GemStone processes list as Killed when I run the gslist command. I’m going to have to dig up my notes on how to get the whole system up and running again, as it has been running nonstop for something like two years, IIRC. As I’m no longer actively working on this project, I’m at a little bit of a loss as to all the things I have to do, LOL. >> >> Anything I should be aware of when trying to bring it all back up? I have to bring up nginx, FastCGI, and all the rest. It should just restore to the last commit point in the logs, I assume. >> >> Regards, >> >> Larry >> _______________________________________________ >> Glass mailing list >> [hidden email] >> http://lists.gemtalksystems.com/mailman/listinfo/glass > _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
I’m struggling with getting the squeak image to display windows on my laptop. I upgraded XQuartz for X11 support under Maverick, and I ran ‘host +’ to allow any host to display windows on my machine. I set ‘export DISPLAY=myipaddress:0’ on the Amazon instance and bring up Squeak, but no windows display on my machine! This has always worked for me before, so I’m puzzled as to what is wrong. Squeak is running fine in the background. If I do get the Pharo windows to display again, do I have to start anything from the image to get Seaside up and running? I start FastCGI Gems from a script. Any thoughts? Larry On Sep 16, 2014, at 2:46 AM, Lawrence Kellogg via Glass <[hidden email]> wrote: > Well, it looks like I can start and stop Gemstone without any problem, as well as netldi, but I’m having trouble getting nginx to talk to my fastcgi gems. Perhaps it has something to do with my A record from Register.com, I’m not sure. I checked the IP address of my running Amazon instance, and I see that I have an elastic IP address allocated, which is in the A record on Register. Register sticks some other ip address in the first line of my A record, something I have noticed in the past. I removed that other ip address, but changes take a while to propagate. > > Strangely, while the ip address seems fine in the nginx error log, the hostname is wrong, it somehow to think it is a different server than the one I’m accessing. > > upstream timed out (110: Connection timed out) while connecting to upstream, client: zz.zz.zz.zz, server: , request: "GET / HTTP/1.1", upstream: "fastcgi://xx.xx.xx.xx:9001", host: “www.practjour.com" > > (should be www.practicemusic.com) I also own www.practjour.com. > > I wrote a little supervise job to bring up the fast cgi gems automatically if they ever fall down. that seems to work. I changed my nginx.conf file to make sure it points to the ip address of my server, plus :9001, :9002, and :9003 for the gems. > > So, I’m kind of stumped at the moment. I’ll work on it some more later. > > Larry > > > On Sep 15, 2014, at 9:59 PM, James Foster <[hidden email]> wrote: > >> The “Killed” status can be cleared by using ‘gslist -c’ to “clear” the lock files (typically in /opt/gemstone/locks). And, yes, the GemStone system will restore to the last commit point. As to the other processes, I don’t have much to contribute. >> >> James >> >> On Sep 15, 2014, at 6:42 PM, Lawrence Kellogg via Glass <[hidden email]> wrote: >> >>> Well, I knew it had to happen eventually. It looks like Amazon took down my instance and did some maintenance on it, so all of my GemStone processes list as Killed when I run the gslist command. I’m going to have to dig up my notes on how to get the whole system up and running again, as it has been running nonstop for something like two years, IIRC. As I’m no longer actively working on this project, I’m at a little bit of a loss as to all the things I have to do, LOL. >>> >>> Anything I should be aware of when trying to bring it all back up? I have to bring up nginx, FastCGI, and all the rest. It should just restore to the last commit point in the logs, I assume. >>> >>> Regards, >>> >>> Larry >>> _______________________________________________ >>> Glass mailing list >>> [hidden email] >>> http://lists.gemtalksystems.com/mailman/listinfo/glass >> > > _______________________________________________ > Glass mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/glass _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
On Tue, Sep 16, 2014 at 7:34 AM, Lawrence Kellogg via Glass <[hidden email]> wrote:
Typically I use `ssh -X` (port forwarding) to connect to the remote machine ... then starting squeak on the remote machine will display locally ...
The script should be fine I would think ... unless you have changed things in your image... Dale _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
Administrator
|
And if you connect via Putty, for example, make sure the connection enables port forwarding. In Putty, this is found in the Connection/SSH/X11 part of the configuration tree. p.s. We really need to do something about the stability of GemStone, so that people can exercise restart scenarios more often. With multiple years going by between restarts, it is just too difficult to remember all the details! |
Re: [Glass] Need to restart GLASS...
|
|
On Sep 16, 2014, at 12:35 PM, Richard Sargent via Glass <[hidden email]> wrote: > Dale wrote >> Typically I use `ssh -X` (port forwarding) to connect to the remote >> machine >> ... then starting squeak on the remote machine will display locally ... > > And if you connect via Putty, for example, make sure the connection enables > port forwarding. In Putty, this is found in the Connection/SSH/X11 part of > the configuration tree. > Ok, some progress, I got the Pharo image to come up through X11. Thanks to John McIntosh for suggesting that I needed to set the port forwarding for X11Forwarding and AllowTcpForwarding in /etc/ssh/sshd_config on my Amazon instance. Not sure why I never ran into this in the past. X11Forwarding was yes but AllowTcpForwarding was no. I’m still stuck though, for some reason the nginx through FastCGI chain is not quite working. Do I need to reinitialize my login task? I tried this: | application | application := (WAAdmin register: self asApplicationAt: 'PracticeJournalLoginTask') exceptionHandler: WAGemStoneProductionErrorHandler. WAAdmin defaultDispatcher defaultName: 'PracticeJournalLoginTask'. application preferenceAt: #sessionClass put: PracticeJournalSession but it didn’t make any difference. How does the system know to go to my login task? Is there something I need to put in my A record on Register.com? Boy, I used to know this stuff but I have forgotten. I’m only running one FastCGI process, on port 9001. That never seemed to bother things before.. I see the Topaz script running when I do a ps -ef. > > > p.s. > We really need to do something about the stability of GemStone, so that > people can exercise restart scenarios more often. With multiple years going > by between restarts, it is just too difficult to remember all the details! > > Yes, two years is a long time, Gemstone is just too reliable. Please crash it randomly every week so that I remember how to pick it back up. ;-) I should have captured a screen movie of one of my restarts. I wrote down notes but must have left out some of the obvious details! Larry > > > -- > View this message in context: http://forum.world.st/Glass-GLASS-Need-to-restart-GLASS-tp4778235p4778414.html > Sent from the GLASS mailing list archive at Nabble.com. > _______________________________________________ > Glass mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/glass _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
On Tue, Sep 16, 2014 at 11:17 AM, Lawrence Kellogg via Glass <[hidden email]> wrote:
I think we need to see some detailed information about what "not working" means: -- http error -- corresponding nginx log Then we might be able to figure things out a bit more ... Dale _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
On Sep 16, 2014, at 2:31 PM, Dale Henrichs <[hidden email]> wrote:
Ok, from nginx it looks like I’m getting time outs 2014/09/16 18:55:08 [error] 11070#0: *1 upstream timed out (110: Connection timed out) while connecting to upstream, client: 68.175.51.239, server: , request: "GET / HTTP/1.1", upstream: "<a href="fastcgi://23.21.100.123:9001">fastcgi://23.21.100.123:9001", host: “www.practicemusic.com" hmm…GET / A few lines from the nginx access.log 68.175.51.239 - - [16/Sep/2014:18:11:59 +0000] "GET /splash_logo_element.png HTTP/1.1" 200 13551 "http://www.practicemusic.com/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.78.2 (KHTML, like Gecko) Version/7.0.6 Safari/537.78.2" "-" aaa68.175.51.239 - - [16/Sep/2014:18:11:59 +0000] "GET /oneColFixCtrHdr.css HTTP/1.1" 200 3376 "http://www.practicemusic.com/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.78.2 (KHTML, like Gecko) Version/7.0.6 Safari/537.78.2" "-" 220.181.108.176 - - [16/Sep/2014:18:54:48 +0000] "GET / HTTP/1.1" 301 185 "-" "Mozilla/5.0 (compatible; Baiduspider/2.0; +http://www.baidu.com/search/spider.html)" "-" 68.175.51.239 - - [16/Sep/2014:18:56:08 +0000] "GET / HTTP/1.1" 504 3019 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.78.2 (KHTML, like Gecko) Version/7.0.6 Safari/537.78.2" "-" 68.175.51.239 - - [16/Sep/2014:18:56:08 +0000] "GET /oneColFixCtrHdr.css HTTP/1.1" 200 3376 "http://www.practicemusic.com/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.78.2 (KHTML, like Gecko) Version/7.0.6 Safari/537.78.2" "-" 68.175.51.239 - - [16/Sep/2014:18:56:08 +0000] "GET /splash_logo_element.png HTTP/1.1" 200 13551 "http://www.practicemusic.com/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.78.2 (KHTML, like Gecko) Version/7.0.6 Safari/537.78.2" "-" 123.125.71.102 - - [16/Sep/2014:18:56:37 +0000] "GET / HTTP/1.1" 301 185 "-" "Mozilla/5.0 (compatible; Baiduspider/2.0; +http://www.baidu.com/search/spider.html)" "-" I am seeing this error in the log for FastCGI 9001, up through 9/15, from !!!!------- Unreportable ERROR Encountered: 2014-09-15T08:54:41.82931900024414-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.----------- !!!!------- Unreportable ERROR Encountered: 2014-09-15T08:56:41.87794589996338-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.----------- !!!!------- Unreportable ERROR Encountered: 2014-09-15T08:58:41.9195671081543-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.----------- !!!!------- Unreportable ERROR Encountered: 2014-09-15T09:00:41.96276998519897-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.----------- !!!!------- Unreportable ERROR Encountered: 2014-09-15T09:02:41.99716711044312-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.----------- --- 09/15/14 16:04:24.459 UTC :stone oobSocket read failed 19 recv(5,0x7fb6be96ee5f,1,0) failed with errno=4,EINTR, Interrupted system call (signal has aborted a sys call) which appears to come from this walk back: InterpreterError 2008: A method has been invoked in the abstract superclass <MultiByteString> that was designed to have been overridden in a concrete subclass. Selector: <#'new'>. 1 GRGemStonePlatform >> logError:title:shouldCommit: @2 line 4 [GsMethod OOP 175180033] 2 GRGemStonePlatform >> logError:title: @2 line 3 [GsMethod OOP 175180289] 3 WAFastCGIAdaptor >> internalServerErrorMessage: @20 line 14 [GsMethod OOP 212443905] 4 ComplexBlock in GRGemStonePlatform >> seasideProcessRequest:adaptor:resultBlock: @29 line 20 [GsMethod OOP 175179521] 5 ExceptionHandler >> caughtExceptionWithAction: @5 line 4 [GsMethod OOP 10065153] 6 ComplexBlock in ExceptionHandler >> caughtEx:number:cat:args:handler: @12 line 13 [GsMethod OOP 10064641] 7 ComplexBlock in ExecutableBlock >> ensure: @4 line 11 [GsMethod OOP 2304001] 8 ComplexBlock in ExecutableBlock >> ensure: @6 line 11 [GsMethod OOP 2304001] 9 ExceptionHandler >> caughtEx:number:cat:args:handler: @16 line 14 [GsMethod OOP 10064641] 10 ComplexBlock in ExceptionHandler >> try:on:do: @10 line 12 [GsMethod OOP 10062081] 11 Exception >> _signal:number:args: @2 line 7 [GsMethod OOP 2327041] 12 System class >> signal:args:signalDictionary: @5 line 13 [GsMethod OOP 3241473] 13 Object >> _error:args: @7 line 10 [GsMethod OOP 1926401] 14 Object >> subclassResponsibility: @2 line 7 [GsMethod OOP 1923841] 15 MultiByteString class >> new @1 line 3 [GsMethod OOP 7129345] 16 MultiByteString >> encodeUsing: @2 line 11 [GsMethod OOP 7135745] 17 WAEncoder >> nextPutAll: @2 line 5 [GsMethod OOP 177867009] 18 GRNullCodecStream >> nextPutAll: @4 line 3 [GsMethod OOP 119043329] 19 ComplexBlock in WAUrl >> encodePathOn: @7 line 5 [GsMethod OOP 176875521] 20 SequenceableCollection >> keysAndValuesDo: @5 line 4 [GsMethod OOP 19120897] This error goes back to September 9th, so the server has been running for weeks with this problem…. Not sure this is helpful. I’ll keep digging. Thanks, Dale and John…. Larry ~
_______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Re: [Glass] Need to restart GLASS...
|
|
The !!!!------- Unreportable ERROR Encountered: 2014-09-15T09:00:41.96276998519897-07:00InterpreterError 2089: An index range was specified for a sequenceable collection with the starting index <1> greater than the ending index <0>.------ errors are probably due to some process (monit?) hitting the url[1].... Regarding the other problem, what version of GemStone are you using? The implementation of MultiByteString>>encodeUsing: is suspect ... Dale On Tue, Sep 16, 2014 at 12:10 PM, Lawrence Kellogg <[hidden email]> wrote:
_______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
«
Return to GLASS
|
1 view|%1 views
| Free forum by Nabble | Edit this page |