Hello,
I'm having an extrange error while tryng to start a repository. Here is the important part of "repository.log": * ======================================================================== Now starting GemStone monitor. Write to /proc/6578/oom_score_adj failed with EACCES , linux user does not have CAP_SYS_RESOURCE No server process protection from OOM killer _____________________________________________________________________________ | SESSION CONFIGURATION: The maximum number of concurrent sessions is 40. | |_____________________________________________________________________________| Attaching the Shared Cache using Stone name: Newstone WARNING: CONCURRENCY_MODE is not a known GemStone option. Started free frame page server, processId 6587 Successfully started 1 free frame page servers. ------------------------------------------------------- Summary of Configured Transaction Logs Directory 0: configured name $repositoryDir/tranlogs expanded name /mnt/disk3/New/tranlogs/ configuredSize 100 MB Directory 1: configured name $repositoryDir/tranlogs expanded name /mnt/disk3/New/tranlogs/ configuredSize 100 MB ------------------------------------------------------- Started AIO page server, processId 6590 --- 09/21/17 11:54:22 -03 --- Extent #0 ----------- Filename = !#dbf!/mnt/disk3/New/data/extent_01.dbf Maximum size = NONE File size = 70 Mbytes = 4480 pages Space available = 48 Mbytes = 3109 pages Totals ------ Repository Size = 70 Mbytes = 4480 pages Free Space = 48 Mbytes = 3109 pages --------------------------------------------------- In extent 0, maxSize is inconsistent. Repository was not shutdown cleanly, recovery needed. Starting recovery of extent sizes changing the maximum size from UNLIMITED MB to 3500 MB for extent: !#dbf!/mnt/disk3/New/data/extent_01.dbf Finished recovery of extent size changes Repository startup statistics: Pages Need Reclaiming =1 Free Oops=9144 Oop Number High Water Mark=130800 Possible Dead Objects=0 Dead Objects=0 Epoch Transaction Count=0 Epoch New Objects Union=0 Epoch Written Objects Union=0 Epoch DependencyMap Objects Union=0 Repository startup is from checkpoint = (fileId 0, blockId 32305) SearchForMostRecentLog did not find any tranlogs :: (wildcard) found in listening addresses, ignoring other addresses created listening socket for :: on :: port 39755 Opened page manager thread log /mnt/disk3/New/logs/repository_6578pagemanager.log Creating a new transaction log. directoryId 0, fileId 1, filename = /mnt/disk3/New/tranlogs/tranlog1.dbf --- 09/21/17 11:54:24 -03 --- Starting AF_INET reconnect to AIO pgsvrs AioServerReconnect pgsvrPid 6590 old fd 12 new fd -1 reconnect failed(B) for AIO pgsvr pid 6590 --- 09/21/17 11:56:25 -03 --- RDbfAioServersReconnect failed Terminating stone. * As I understand, the problem is related to de page server manager, but I tried with different configurations on the machine but couldn't resolve this problem. This problem started recently, after shutdown of the machine. I am pretty sure that there is some configuration that I'm missing, but I didn't find the problem yet. The OS is Linux CentOS 6 and I'm using gemstone 3.2.8.1. Could you help me please? Which configurations should i check? Thanks in advance! Ezequiel. -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
On Fri, Sep 22, 2017 at 4:23 PM, brianstone via GemStone-Smalltalk <[hidden email]> wrote: Hello, Yeah, the issue does not seem to be in the stone process itself. You should send the rest of the logs. Maybe there is something useful in XXXpcmon.log This problem started recently, after shutdown of the machine. _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
It's true, sorry.
Here is *pcmon.log: * _____________________________________________________________________________ | GemStone Child Task | | | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 6580 DATE: 09/21/17 11:54:20 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | | COMMAND: /usr/local/gemstone643281/sys/startshrpcmon Newstone~7663a27bab8c7a96 | 312500 51 1 5000 7000 1 1900 0 432 0 60 | |_____________________________________________________________________________| startshrpcmon.sh GEMSTONE: /usr/local/gemstone643281 SharedPageCache Monitor Summary: the hostname is: SRVR23 GEMSTONE is: "/usr/local/gemstone643281" shrpcmon arguments are: SharedPageCache Monitor name Newstone~7663a27bab8c7a96. Number of pages 312500. Number of processes 51. Cache locked in memory 1. SpinLockCount 5000. TargetFreeFrameCount 7000. Number of free frame servers 1. Number of shared counters 1900. Port number 0. Permissions 432. Large Page Policy 0. Stone/cachepgsvr connect timeout 60. _____________________________________________________________________________ | GemStone/S64 Object-Oriented Data Management System | | Copyright (C) GemTalk Systems 1986-2015 | | All rights reserved. | +-----------------------------------------------------------------------------+ | PROGRAM: SHRPCMON, GemStone SharedPageCache Monitor | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 6580 DATE: 09/21/17 11:54:20 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | |_____________________________________________________________________________| Cache config is 312500 pages = 4882MB, total is 5232MB, overhead 7% of configured size [Info]: Building the free frame list using 1 cache region. WARNING: Shared cache could not be locked in memory because the memory available is insufficient. (ENOMEM) Shared memory region number 1 has id = 360452 The shared semaphore array has id = 327682 --- 09/21/17 11:54:21.258 -03 cache creation took 1 seconds created listening socket for localhost6 on ::1 port 38772 synthesizing 127.0.0.1 , attempting listen on IPv4 localhost created listening socket for localhost on ::ffff:127.0.0.1 port 38772 --- 09/21/17 11:54:21.273 -03 Listening for clients on port 38772 Write to /proc/6580/oom_score_adj failed with EACCES , linux user does not have CAP_SYS_RESOURCE No server process protection from OOM killer timeout for stone/cache pgsvr connect is 20 seconds [09/21/17 11:56:25.819 -03]: Client died: Slot 1, PID 6578, LostOtFlags 0, sessionId 0 Name Newstone --- 09/21/17 11:56:25.820 -03 Starting shrpcmonitor shutdown --- 09/21/17 11:56:25.820 -03 Waiting for crashed slot recovery thread to shutdown...Done. --- 09/21/17 11:56:25.821 -03 Waiting for clean slot recovery thread to shutdown...Done. --- 09/21/17 11:56:25.821 -03 Waiting for stats thread to shutdown...Done. --- 09/21/17 11:56:25.890 -03 All threads have stopped. The Shared Page Cache Monitor is shutting down. * And here is *pagemanager.log: * _____________________________________________________________________________ | Stone Page Manager Thread | | | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 6578 DATE: 09/21/17 11:54:22 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | |_____________________________________________________________________________| [Info]: Initial Page Manager Settings: StnPageMgrRemoveMinPages: 40 StnPageMgrRemoveMaxPages: 16384 PageMgrCompressionEnabled: FALSE PageMgrPrintTimeoutThreshold: 5 StnRemoteCachePgsvrTimeout: 0 RemoteSharedPageCacheMax: 255 * I'll be waiting for your comments. Thanks -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
That's weird...no error in pcmon either. Can you check and send us the logs of pgsvr ?? Thanks On Mon, Sep 25, 2017 at 11:41 AM, brianstone via GemStone-Smalltalk <[hidden email]> wrote: It's true, sorry. _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
In reply to this post by Gemstone/S mailing list
In particular, see the following from the stone log:
--- 09/21/17 11:54:24 -03 --- Starting AF_INET reconnect to AIO pgsvrs AioServerReconnect pgsvrPid 6590 old fd 12 new fd -1 reconnect failed(B) for AIO pgsvr pid 6590 Is there a log file for PID 6590?
_______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi James,
There was not a log for that PID. I was reading the SysAdminGuide and there says that the log is deleted automatically in case of normal exit. In this case is not a normal exit but anyway log files are deleted. So, I modified the file runpgsrvrmain to keep log files (uncommenting "export GEMSTONE_KEEP_LOG=1") and tried to start again the repository. Here is the results: *An extract from repository.log* Starting AF_INET reconnect to AIO pgsvrs Recovery took 0.028 seconds Waiting for Recovery Reader thread to stop AioServerReconnect pgsvrPid 7940 old fd 12 new fd -1 reconnect failed(B) for AIO pgsvr pid 7940 --- 09/25/17 12:59:18 -03 --- RDbfAioServersReconnect failed Terminating stone. *Log file for PID 7940 in file named "repository_7940pgsvraio.log"* _____________________________________________________________________________ | GemStone Child Task | | | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 7940 DATE: 09/25/17 12:57:17 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | | COMMAND: /usr/local/gemstone643281/sys/runpgsvrmain TCP 13 90 | |_____________________________________________________________________________| runpgsvr[Info]: Description of arguments: the hostname is: SRVR23 GEMSTONE is: /usr/local/gemstone643281 pgsvr arguments are: TCP 13 90 _____________________________________________________________________________ | GemStone/S64 Object-Oriented Data Management System | | Copyright (C) GemTalk Systems 1986-2015 | | All rights reserved. | +-----------------------------------------------------------------------------+ | PROGRAM: PGSVRSHR, GemStone Networked DBF I/O Service (shared library) | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 7940 DATE: 09/25/2017 12:57:17 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | |_____________________________________________________________________________| command line is: /usr/local/gemstone643281/sys/pgsvrmain TCP 13 90 The hostname is SRVR23 createNetConnection: SocketFamily_UNIX Network connection has been inherited. Entering Service Loop [Info]: ClientPid: 7928 [Info]: Client SessionId: -2 [Info]: Client Host: SRVR23 [Info]: My cache slot: 3 [Info]: My cache name: AioPgsvr3 Write to /proc/7940/oom_score_adj failed with EACCES , linux user does not have CAP_SYS_RESOURCE No server process protection from OOM killer --- 09/25/2017 12:59:18.753 -03 Connection failure detected: --- 09/25/2017 12:59:18.753 -03 entering pgsShrExit mainThread Detaching cache.--- 09/25/2017 12:59:18.753 -03 [Info]: Detaching Shared Page Cache. --- 09/25/2017 12:59:18.791 -03 mainThread: pgsShrExit with status: 0 *Additionally heres is the content of another file named "repository_7937pgsvrff.log"* _____________________________________________________________________________ | GemStone Child Task | | | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 7937 DATE: 09/25/17 12:57:17 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | | COMMAND: /usr/local/gemstone643281/sys/runpgsvrmain Newstone~7663a27bab8c7a96 | 0 1 -1 TCP 10 90 | |_____________________________________________________________________________| runpgsvr[Info]: Description of arguments: the hostname is: SRVR23 GEMSTONE is: /usr/local/gemstone643281 pgsvr arguments are: Newstone~7663a27bab8c7a96 0 1 -1 TCP 10 90 _____________________________________________________________________________ | GemStone/S64 Object-Oriented Data Management System | | Copyright (C) GemTalk Systems 1986-2015 | | All rights reserved. | +-----------------------------------------------------------------------------+ | PROGRAM: PGSVRSHR, GemStone Networked DBF I/O Service (shared library) | | VERSION: 3.2.8.1, Fri Aug 28 08:43:23 2015 | | BUILD: gss64_3_2_x_branch-37291 | | BUILT FOR: x86-64 (Linux) | | MODE: 64 bit | | RUNNING ON: 12-CPU SRVR23 x86_64 (Linux 2.6.32-642.15.1.el6.x86_64 #1 SMP Fri | Feb 24 14:31:22 UTC 2017) 32059MB | | PROCESS ID: 7937 DATE: 09/25/2017 12:57:17 -03 | | USER IDS: REAL=gemst643281 (1045) EFFECTIVE=gemst643281 (1045) LOGIN=gemst643281 | (1045) | |_____________________________________________________________________________| command line is: /usr/local/gemstone643281/sys/pgsvrmain Newstone~7663a27bab8c7a96 0 1 -1 TCP 10 90 The hostname is SRVR23 Write to /proc/7937/oom_score_adj failed with EACCES , linux user does not have CAP_SYS_RESOURCE No server process protection from OOM killer Free Frame Page Server startup was successful. Target Free Frame Limit is 7000 Entering Free List Service Loop. createNetConnection: SocketFamily_UNIX Network connection has been inherited. Entering Service Loop [Info]: ClientPid: 7928 --- 09/25/2017 12:59:18.753 -03 Connection failure detected: --- 09/25/2017 12:59:18.753 -03 entering pgsShrExit mainThread Detaching cache.--- 09/25/2017 12:59:18.753 -03 [Info]: Detaching Shared Page Cache. --- 09/25/2017 12:59:18.753 -03 mainThread: pgsShrExit with status: 0 -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi Ezequiel,
From the initial stone log we see that a listening socket was created, the AIO pgsvrs were started, but the reconnect failed:
Note the 121-second gap and the ‘-1’ value for the new fd (file descriptor or socket ID), indicating an error. From the later page server log we see the process start date/time and the connection failure date/time:
Note again a 121-second gap. As part of stone startup, the stone spawns additional processes and waits for them to connect back to itself. The system waits 120 seconds for the connection to complete and then reports a failure. So, the page server is unable to open a socket to the stone (that is, a networking problem). Is this machine heavily overloaded, such that a new process would fail to run in two minutes? Have you every run GemStone successfully on this host? What sort of internal firewall does it have? During the two-minute delay in the startup, it would be interesting to see if you can communicate with the stone using another process. From another shell try the following $ gslist -cvl If the shell doesn’t recognize gslist, then you don’t have $GEMSTONE/bin in your path. If it doesn’t find any servers, then you haven’t started the NetLDI process yet and it would be good to do so (if only to show that gslist is working!). If it finds the stone and reports that the stone is in ‘startup’ mode, then gslist has successfully communicated with the stone, suggesting that internal networking is not completely broken. If gslist recognizes that there is a stone, but can’t communicate with it, then we’ve confirmed an internal networking problem. To further debug networking problems, export GEMSTONE_SOCKET_DEBUG=1 (as you did with GEMSTONE_KEEP_LOG; good job reading the manual!), and then provide full logs on the stone and on a failing page server process. James Foster
_______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
James,
Here are all the log files obtained using "GEMSTONE_SOCKET_DEBUG=1" logs.zip <http://forum.world.st/file/t370666/logs.zip> On the other hand, during the startup I've tried many times with gslist to see the result. Using gslist-l gslist -l Status Version Owner Pid Port Started Type Name ------- --------- --------- ----- ----- ------------ ------ ---- exists 3.2.8.1 gemst643281 9878 37052 Sep 25 15:48 cache Newstone~7663a27bab8c7a96 exists 3.2.8.1 gemst643281 9876 38044 Sep 25 15:48 Stone Newstone exists 3.2.8.1 gemst643281 7912 50387 Sep 25 12:56 Netldi gs64ldi-3281 Using gslist -clv gslist -clv Status Version Owner Pid Port Started Type Name ------- --------- --------- ----- ----- ------------ ------ ---- OK 3.2.8.1 gemst643281 9878 37052 Sep 25 15:48 cache Newstone~7663a27bab8c7a96 frozen 3.2.8.1 gemst643281 9876 38044 Sep 25 15:48 Stone Newstone OK 3.2.8.1 gemst643281 7912 50387 Sep 25 12:56 Netldi gs64ldi-3281 Based on the logs, I think that you are right and is a networking problem. Regarding your questions: This server belongs to the company which I work for. It's often used to keep some internal web sites running during a short period, for example for internal contests. Since I'm the administrator of the server and I have full control over it, some time ago I decided to install and test Gemstone. Currently is not running any site, and is unused, so the server is not overloaded. Recently this machine have been moved to a different network which have a more restrictive firewall, but I requested to enable traffic through the netldi port (I thought it will be enough). Summarizing, yes, I was able to run Gemstone before the network change, but I thought that enabling netldi port should be enough. What are the networking requisites needed to start Gemstone? Hope this info can help you to deveal the problem. Ezequiel -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
On Mon, Sep 25, 2017 at 4:41 PM, brianstone via GemStone-Smalltalk <[hidden email]> wrote: James, I think that in the past (not sure which versions), GemStone would use 1 or 2 more ports aside from the netldi one (netldi+1 and netldi+2 or something like that). I think in latest Gemstone, netldi was able to use only one part for all traffics. I suspect trying to start netldi with -p option, like this: $GEMSTONE/bin/startnetldi -P 50377 -p50378:50379 -g -a $USER may help (of course, be sure to open 50378 and 50379 in the firewall). at least not as a definitive workaround but for troubleshooting. Cheers, What are the networking requisites needed to start Gemstone? _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Mariano,
You are right about the NetLDI ports and changes with respect to a RPC client communicating with the NetLDI and the new Gem, but Ezequiel is dealing with an AIO page server talking to the stone, and this is a different port. James
_______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
In reply to this post by Gemstone/S mailing list
Ezequiel,
In general, a firewall will restrict communications between hosts, but you can also configure a firewall to restrict communications between processes on a single host. If your machine is configured to limit communications between processes that are started by the same user, then you need to enable some additional ports. By default, the various processes listen on a random port (as assigned by the OS). You can use various configurations to specify the actual port used. This is typically more important with respect to the NetLDI process, since it is called from external processes, but it can be an issue for other processes as well. In the System Administration Guide, search for ‘WELL_KNOWN’ and then configure your system to use well-known ports for the Stone and the Shared Page Cache Monitor. Then enable those ports in your internal firewall and try starting the system again and let us know the result. James Foster > On Sep 25, 2017, at 12:41 PM, brianstone via GemStone-Smalltalk <[hidden email]> wrote: > > James, > > Here are all the log files obtained using "GEMSTONE_SOCKET_DEBUG=1" > > logs.zip <http://forum.world.st/file/t370666/logs.zip> > > On the other hand, during the startup I've tried many times with gslist to > see the result. > Using gslist-l > > gslist -l > Status Version Owner Pid Port Started Type Name > ------- --------- --------- ----- ----- ------------ ------ ---- > exists 3.2.8.1 gemst643281 9878 37052 Sep 25 15:48 cache > Newstone~7663a27bab8c7a96 > exists 3.2.8.1 gemst643281 9876 38044 Sep 25 15:48 Stone > Newstone > exists 3.2.8.1 gemst643281 7912 50387 Sep 25 12:56 Netldi > gs64ldi-3281 > > Using gslist -clv > > gslist -clv > Status Version Owner Pid Port Started Type Name > ------- --------- --------- ----- ----- ------------ ------ ---- > OK 3.2.8.1 gemst643281 9878 37052 Sep 25 15:48 cache > Newstone~7663a27bab8c7a96 > frozen 3.2.8.1 gemst643281 9876 38044 Sep 25 15:48 Stone > Newstone > OK 3.2.8.1 gemst643281 7912 50387 Sep 25 12:56 Netldi > gs64ldi-3281 > > > Based on the logs, I think that you are right and is a networking problem. > > Regarding your questions: > > This server belongs to the company which I work for. It's often used to keep > some internal web sites running during a short period, for example for > internal contests. Since I'm the administrator of the server and I have full > control over it, some time ago I decided to install and test Gemstone. > Currently is not running any site, and is unused, so the server is not > overloaded. > Recently this machine have been moved to a different network which have a > more restrictive firewall, but I requested to enable traffic through the > netldi port (I thought it will be enough). > > Summarizing, yes, I was able to run Gemstone before the network change, but > I thought that enabling netldi port should be enough. > > What are the networking requisites needed to start Gemstone? > > Hope this info can help you to deveal the problem. > > Ezequiel > > > > > -- > Sent from: http://forum.world.st/Gemstone-S-f1461796.html > _______________________________________________ > GemStone-Smalltalk mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi,
I followed James's advice and configured the ports for the Stone and the Shared Page Cache Monitor. I'm using this configuration: STN_WELL_KNOWN_PORT_NUMBER = 51001; SHR_WELL_KNOWN_PORT_NUMBER = 51002; Regarding an internall firewall, I tested by deactivating any firewall in the machine but anyway the problem still there. On my last test I have activated again iptables and configured to allow traffic through these ports(51001 and 51002). After that, I tried again to start Gemstone, and during the startup I tested the following commands: *netstat -uptona | grep 51001* tcp 0 0 :::51001 :::* LISTEN 30145/stoned off (0.00/0/0) tcp 160 0 ::1:51001 ::1:47588 ESTABLISHED 30145/stoned off (0.00/0/0) tcp 0 0 ::1:47588 ::1:51001 ESTABLISHED 30157/pgsvrmain off (0.00/0/0) *netstat -uptona | grep 51002* tcp 0 0 ::ffff:127.0.0.1:51002 :::* LISTEN 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:51002 :::* LISTEN 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:45972 ::1:51002 ESTABLISHED 30157/pgsvrmain off (0.00/0/0) tcp 0 0 ::1:45970 ::1:51002 ESTABLISHED 30154/pgsvrmain off (0.00/0/0) tcp 0 0 ::1:45968 ::1:51002 ESTABLISHED 30145/stoned off (0.00/0/0) tcp 0 0 ::1:51002 ::1:45974 ESTABLISHED 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:51002 ::1:45970 ESTABLISHED 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:51002 ::1:45968 ESTABLISHED 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:51002 ::1:45972 ESTABLISHED 30147/shrpcmonitor off (0.00/0/0) tcp 0 0 ::1:45974 ::1:51002 ESTABLISHED 30145/stoned off (0.00/0/0) *nmap -p1-52000 localhost* Starting Nmap 5.51 ( http://nmap.org ) at 2017-09-28 11:49 -03 Nmap scan report for localhost (127.0.0.1) Host is up (0.0000040s latency). rDNS record for 127.0.0.1: SRVR23 Not shown: 51992 closed ports PORT STATE SERVICE 22/tcp open ssh 25/tcp open smtp 111/tcp open rpcbind 631/tcp open ipp 5989/tcp open wbem-https 27017/tcp open unknown 51001/tcp open unknown 51002/tcp open unknown Nmap done: 1 IP address (1 host up) scanned in 1.04 seconds In the result of 'netstat' shows that there is an entablished connection. For me, that is a clearly proof that the connection is not a problem. I'm right? I'm suspecting about a configuration with file descriptors. Can it be a problem? Thanks in advance. Ezequiel -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi Ezequiel, It appears from your observations below and the earlier log files (with extra socket logging) that you are right that at least some connections are succeeding. Thus, it is probably not a firewall issue. After further review of the logs it does seem that there is a problem with how the AIO page server is connecting to the stone. In one of the logs we see a response of SOCK_EINPROGRESS but there is never a full connection. Without looking deeper, we aren’t really sure why the failure happens. Is there a reason you are using version 3.2.8.1 and not something more recent? That was released in August 2015, so is over two years old. Could you try starting a stone for the most recent version, say 3.3.6 or even 3.4.0-Alpha5 (see http://downloads.gemtalksystems.com/pub/GemStone64/)? I think we could get more traction in investigating this on a newer version. Regards, James Foster
_______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi James,
I finally found the problem. There was missing a line in the hosts file. I was looking at this part on the pcmon.log *created listening socket for localhost6 on ::1 port 41793 synthesizing 127.0.0.1 , attempting listen on IPv4 localhost created listening socket for localhost on ::ffff:127.0.0.1 port 41793* Since I'm not an expert on this, I found strange that there is created a socket in localhost6 and one in localhost (as I understand IPV4). So, after a while googling, I noticed that in general, other hosts files in Linux systems, have this line: *::1 localhost localhost.localdomain localhost6 localhost6.localdomain6* So, I added that line to my host file and then Gemstone started succesfully :-) The reason of this behavior seems to be the fact that now I don't have a DNS in the new network of the server or something like that. You are free to add any comment you consider relevant, Thanks you for all your help James, Ezequiel -- Sent from: http://forum.world.st/Gemstone-S-f1461796.html _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Hi Ezequiel,
I’m glad you got the problem resolved. I’ll report this to the engineering team to see if the failure can be made more explicit. James > On Oct 2, 2017, at 7:44 AM, brianstone via GemStone-Smalltalk <[hidden email]> wrote: > > Hi James, > > I finally found the problem. > > There was missing a line in the hosts file. > > I was looking at this part on the pcmon.log > > *created listening socket for localhost6 on ::1 port 41793 > synthesizing 127.0.0.1 , attempting listen on IPv4 localhost > created listening socket for localhost on ::ffff:127.0.0.1 port 41793* > > Since I'm not an expert on this, I found strange that there is created a > socket in localhost6 and one in localhost (as I understand IPV4). > > So, after a while googling, I noticed that in general, other hosts files in > Linux systems, have this line: > *::1 localhost localhost.localdomain localhost6 > localhost6.localdomain6* > > So, I added that line to my host file and then Gemstone started succesfully > :-) > > The reason of this behavior seems to be the fact that now I don't have a DNS > in the new network of the server or something like that. > > You are free to add any comment you consider relevant, > > Thanks you for all your help James, > > Ezequiel > > > > > -- > Sent from: http://forum.world.st/Gemstone-S-f1461796.html > _______________________________________________ > GemStone-Smalltalk mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk _______________________________________________ GemStone-Smalltalk mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk |
Free forum by Nabble | Edit this page |