X-Git-Url: http://scm.dxcluster.org/gitweb/gitweb.cgi?a=blobdiff_plain;ds=inline;f=html%2Fadminmanual-8.html;h=be22c7f6db55bb5060c9e524ed30154bbf7b3d6c;hb=69ffad66605714e1b188f52c2908c9c9c496772a;hp=4d5450a1162215eed59f1ad2c55b1ed435a86c55;hpb=e90334372dd94c9deca9855be9ac7527a0da7870;p=spider.git diff --git a/html/adminmanual-8.html b/html/adminmanual-8.html index 4d5450a1..be22c7f6 100644 --- a/html/adminmanual-8.html +++ b/html/adminmanual-8.html @@ -2,540 +2,108 @@
-Most maintenance tasks are automatic but there are some commands that are useful for a sysop. These are listed below in alphabetical order. The number in brackets following the command name is the permissions level needed to use the command. +
CVS stands for "Concurrent Versions System" and the CVS for DXSpider is held +at +Sourceforge. This means +that it is possible to update your DXSpider installation to the latest +sources by using a few simple commands.
-
announce sysop <text>
-
-
Send an announcement to Sysops only -
-
connect <callsign> Start a connection to another DX Cluster
+
Please be aware that if you update your system using CVS, it is possible that +you could be running code that is very beta and not fully tested. There is +a possibility that it could be unstable.
-
Start a connection process that will culminate in a new connection to the -DX cluster <callsign>. This process creates a new 'client' process which will -use the script in /spider/connect/<callsign> to effect the 'chat' exchange -necessary to traverse the network(s) to logon to the cluster <callsign>. +
I am of course assuming that you have a machine with both DXSpider and +Internet access running.
+
BEFORE YOU EVEN CONSIDER STARTING WITH THIS MAKE A BACKUP OF YOUR +ENTIRE SPIDER TREE!!
-
<node_call> All [<msgno> ...] Mark a message as sent
-
-
When you send messages the fact that you have forwarded it to another node -is remembered so that it isn't sent again. When you have a new partner -node and you add their callsign to your /spider/msg/forward.pl file, all -outstanding non-private messages will be forwarded to them. This may well -be ALL the non-private messages. You can prevent this by using these -commmands:- -
catch GB7DJK all -catch GB7DJK 300 301 302 303 -
and to undo what you have just done:- -
uncatch GB7DJK all -uncatch GB7DJK 300 301 302 303 -
which will arrange for them to be forward candidates again. -
-
dbcreate <name> Create a database entry
+
-dbcreate <name> chain <name> [<name>..] Create a chained database entry
-dbcreate <name> remote <node> Create a remote database entry
Assuming you are connected to the Internet, you need to login to the +CVS repository and then update your Spider source. There are several +steps which are listed below ...
-
DBCREATE allows you to define a database in the system. It doesn't actually -create anything, just defines it. -
The databases that are created are simple DB_File hash databases, they are -therefore already 'indexed'. -
You can define a local database with the first form of the command eg: -
DBCREATE oblast -
You can also chain databases with the addition of the 'chain' keyword. -This will search each database one after the other. A typical example -is: -
DBCREATE sdx_qsl chain sql_ad -
No checking is done to see if the any of the chained databases exist, in -fact it is usually better to do the above statement first then do each of -the chained databases. -
Databases can exist offsite. To define a database that lives on another -node do: -
DBCREATE buckmaster remote gb7dxc -
Remote databases cannot be chained; however, the last database in a -a chain can be a remote database eg: -
DBCREATE qsl chain gb7dxc -
To see what databases have been defined do: -
DBAVAIL (or it will have been aliased to SHOW/COMMAND) -
It would be normal for you to add an entry into your local Aliases file -to allow people to use the 'SHOW/<dbname>' style syntax. So you would -need to add a line like:- +
First login as the user sysop. Next you need to connect to the CVS +repository. You do this with the command below ...
-
- 's' => [
- ..
- ..
- '^sh\w*/buc', 'dbshow buckmaster', 'dbshow',
- ..
- ..
- ],
+cvs -d:pserver:anonymous@cvs.DXSpider.sourceforge.net:/cvsroot/dxspider login
-
-to allow -
SH/BUCK g1tlh -
to work as they may be used to. -
See DBIMPORT for the importing of existing AK1A format data to databases. -See DBSHOW for generic database enquiry -
-
dbimport <dbname> Import AK1A data into a database
+
You will get a password prompt. Simply hit return here and your machine should +return to a normal linux prompt.
-
If you want to import or update data in bulk to a database you can use -this command. It will either create or update entries into an existing -database. For example:- -
DBIMPORT oblast /tmp/OBLAST.FUL -
will import the standard OBLAST database that comes with AK1A into the -oblast database held locally. +
What happens next depends on whether you have an existing installation that +you want to update with the latest and greatest or whether you just want +to see what is there and/or run it on a new machine for testing. +
If you are installing Spider from CVS then change directory to /home/sysop +
If you are wanting to update Spider then cd to /tmp
-
dbremove <dbname> Delete a database
-
-
DBREMOVE will completely remove a database entry and also delete any data -file that is associated with it. -
There is no warning, no comeback, no safety net. -
For example: -
DBREMOVE oblast -
will remove the oblast database from the system and it will also remove -the associated datafile. -
I repeat: -
There is no warning, no comeback, no safety net. -
You have been warned. -
-
debug Set the cluster program into debug mode
+
The next step will create a brand new 'spider' directory in your current +directory.
-
Executing this command will only have an effect if you are running the cluster -in debug mode i.e. -
-
- perl -d cluster.pl
+cvs -z3 -d:pserver:anonymous@cvs.DXSpider.sourceforge.net:/cvsroot/dxspider co spider
-
-It will interrupt the cluster just after the debug command has finished. -
-
Works just like the user command except that sysops can see ALL messages. -
-
disconnect <call> [<call> ...] Disconnect a user or node
+
This command is all on one line.
-
Disconnect any <call> connected locally +
Hopefully your screen should show you downloading files. The -z3 simply compresses +the download to improve speed. +When this has finished, you will have exactly the same as if you had untarred a full +tarball PLUS some extra directories and files that CVS needs to do the magic that +it does.
-
export <msgno> <filename> Export a message to a file
+
Now if you are doing a new installation, that's it. Carry on as if you have +just downloaded and untarred the lastest tarball.
-
Export a message to a file. This command can only be executed on a local -console with a fully privileged user. The file produced will be in a form -ready to be imported back into the cluster by placing it in the import -directory (/spider/msg/import). -
This command cannot overwrite an existing file. This is to provide some -measure of security. Any files written will owned by the same user as the -main cluster, otherwise you can put the new files anywhere the cluster can -access. For example:- -
EXPORT 2345 /tmp/a -
-
forward/opername <call> Send out information on this <call> to all clusters
-
-
This command sends out any information held in the user file which can -be broadcast in PC41 protocol packets. This information is Name, QTH, Location -and Homenode. PC41s are only sent for the information that is available. -
-
init <node call> Re-initialise a link to an AK1A compatible node
-
-
This command attempts to re-initialise a link to a (usually) AK1A node -that has got confused, usually by a protocol loop of some kind. It may -work - but you usually will be better off simply disconnecting it (or -better, if it is a real AK1A node, doing an RCMD <node> DISC/F <your -node>). -
Best of luck - you will need it. -
-
kill <msgno> [<msgno> ...] Remove or erase a message from the system
-
-kill from <call> Remove all messages from a callsign
-kill to <call> Remove all messages to a callsign
-
You can get rid of any message to or originating from your callsign using -this command. You can remove more than one message at a time. -
As a sysop you can kill any message on the system. -
-
kill full <msgno> [<msgno>] Delete a message from the whole cluster
Delete a message (usually a 'bulletin') from the whole cluster system. -
This uses the subject field, so any messages that have exactly the same subject -will be deleted. Beware! -
-
load/aliases Reload the command alias table
-
-
Reload the /spider/cmd/Aliases file after you have editted it. You will need to -do this if you change this file whilst the cluster is running in order for the -changes to take effect. -
-
load/bands Reload the band limits table
-
-
Reload the /spider/data/bands.pl file if you have changed it manually whilst -the cluster is running. -
-
load/cmd_cache Reload the automatic command cache
-
-
Normally, if you change a command file in the cmd or local_cmd tree it will -automatially be picked up by the cluster program. Sometimes it can get confused -if you are doing a lot of moving commands about or delete a command in the -local_cmd tree and want to use the normal one again. Execute this command to -reset everything back to the state it was just after a cluster restart. -
-
load/forward Reload the msg forwarding routing table
-
Reload the /spider/msg/forward.pl file if you have changed it -manually whilst the cluster is running. -
-
load/messages Reload the system messages file
-
-
If you change the /spider/perl/Messages file (usually whilst fiddling/writing ne -commands) you can have them take effect during a cluster session by executing this -command. You need to do this if get something like :- -
unknown message 'xxxx' in lang 'en' -
-
load/prefixes Reload the prefix table
-
-
Reload the /spider/data/prefix_data.pl file if you have changed it manually whilst -the cluster is running. -
-
merge <node> [<no spots>/<no wwv>] Ask for the latest spots and WWV
-
-
MERGE allows you to bring your spot and wwv database up to date. By default -it will request the last 10 spots and 5 WWVs from the node you select. The -node must be connected locally. -
You can request any number of spots or wwv and although they will be appended -to your databases they will not duplicate any that have recently been added -(the last 2 days for spots and last month for WWV data). -
-
msg <cmd> <msgno> [data ...] Alter various message parameters
-
-
Alter message parameters like To, From, Subject, whether private or bulletin -or return receipt (RR) is required or whether to keep this message from timing -out. +
If you want to upgrade your current installation then do this ...
- MSG TO <msgno> <call> - change TO callsign to <call>
- MSG FRom <msgno> <call> - change FROM callsign to <call>
- MSG PRrivate <msgno> - set private flag
- MSG NOPRrivate <msgno> - unset private flag
- MSG RR <msgno> - set RR flag
- MSG NORR <msgno> - unset RR flag
- MSG KEep <msgno> - set the keep flag (message won't be deleted ever)
- MSG NOKEep <msgno> - unset the keep flag
- MSG SUbject <msgno> <new> - change the subject to <new>
- MSG WAittime <msgno> - remove any waitting time for this message
- MSG NOREad <msgno> - mark message as unread
- MSG REad <msgno> - mark message as read
- MSG QUeue - queue any outstanding bulletins
- MSG QUeue 1 - queue any outstanding private messages
+tar cvfz /tmp/s.tgz spider
+cd /
+tar xvfzp /tmp/s.tgz
-You can look at the status of a message by using:- -
STAT/MSG <msgno> -
This will display more information on the message than DIR does. +
This is assuming you downloaded to the /tmp directory of course.
-
pc <call> <text> Send text (eg PC Protocol) to <call>
-
-
Send some arbitrary text to a locally connected callsign. No processing is done on -the text. This command allows you to send PC Protocol to unstick things if problems -arise (messages get stuck etc). eg:- -
pc gb7djk PC33^GB7TLH^GB7DJK^400^ -
You can also use in the same way as a talk command to a connected user but -without any processing, added of "from <blah> to <blah>" or whatever. -
pc G1TLH Try doing that properly!!! +
NOTE: the 'p' on the end of the 'xvfz' is IMPORTANT! It keeps the permissions +correct. YOU WERE LOGGED IN AS THE USER SYSOP WEREN'T YOU????? +
Remember to recompile the C client (cd /spider/src; make)
-
ping <node> Send a ping command to another cluster node
+
At this point the files have been upgraded. You can (usually) restart the cluster +in your own time. However, if you attempt to use any new commands or features +expect it to be fatal! At least your cluster will have been restarted then so it +will be too late to worry about it!
-
This command is used to estimate the quality of the link to another cluster. -The time returned is the length of time taken for a PC51 to go to another -cluster and be returned. -
Any visible cluster node can be PINGed. -
-
rcmd <node call> <cmd> Send a command to another DX cluster
-
-
This command allows you to send nearly any command to another DX Cluster -node that is connected to the system. -
Whether you get any output is dependant on a) whether the other system knows -that the node callsign of this cluster is in fact a node b) whether the -other system is allowing RCMDs from this node and c) whether you have -permission to send this command at all. -
-
read <msgno> Read a message on the system
-
-
As a sysop you may read any message on the system -
-
set/debug <name> Add a debug level to the debug set
-
-
You can remove this level with unset/debug <name> -
-
set/isolate <node call> Isolate a node from the rest of the network
-
-
Connect a node to your system in such a way that you are a full protocol -member of its network and can see all spots on it, but nothing either leaks -out from it nor goes back into from the rest of the nodes connected to you. -
You can potentially connect several nodes in this way. -
You can see which nodes are isolated with the show/isolate (1) command. -
You can remove the isolation with the command unset/isolate. -
-
set/sys_location <lat & long> Set your cluster latitude and longitude
-
-
In order to get accurate headings and such like you must tell the system -what your latitude and longitude is. If you have not yet done a SET/QRA -then this command will set your QRA locator for you. For example:- -
SET/LOCATION 52 22 N 0 57 E -
-
set/lockout <call> Stop a callsign connecting to the cluster
-
-
You can show who is locked out with the show/lockout (9) command. -
To allow the user to connect again, use the command unset/lockout -
-
set/node <call> [<call> ...] Make the callsign an AK1A cluster
-
-
Tell the system that the call(s) are to be treated as AK1A cluster and -fed PC Protocol rather normal user commands. -
From version 1.41 you can also set the following types of cluster -
-
-set/spider -set/dxnet -set/clx -set/arcluster --
To see what your nodes are set to, use the show/nodes command. -
-
set/obscount <count> <node call> Set the 'pump-up' obscelence counter
-
-
From version 1.35 onwards neighbouring nodes are pinged at regular intervals (see -SET/PINGINTERVAL), usually 300 seconds or 5 minutes. There is a 'pump-up' -counter which is decremented on every outgoing ping and then reset to -the 'obscount' value on every incoming ping. The default value of this -parameter is 2. -
What this means is that a neighbouring node will be pinged twice at -(default) 300 second intervals and if no reply has been heard just before -what would be the third attempt, that node is disconnected. -
If a ping is heard then the obscount is reset to the full value. Using -default values, if a node has not responded to a ping within 15 minutes, -it is disconnected. -
-
set/pinginterval <time> <node call> Set the ping time to neighbouring nodes
-
-
As from version 1.35 all neighbouring nodes are pinged at regular intervals -in order to determine the rolling quality of the link and, in future, to -affect routing decisions. The default interval is 300 secs or 5 minutes. -
You can use this command to set a different interval. Please don't. -
But if you do the value you enter is treated as minutes up 60 and seconds -for numbers greater than that. -
This is used also to help determine when a link is down at the far end -(as certain cluster software doesn't always notice), see SET/OBSCOUNT -for more information. -
-
set/privilege <n> <call> [<call> ...] Set the privilege level on a call
-
-
Set the privilege level on a callsign. The privilege levels that pertain -to commands are as default:- +
Now the magic part! From now on when you want to update, simply connect to the +Internet and then, as the user sysop ...
- 0 - normal user
- 1 - allow remote nodes normal user RCMDs
- 5 - various privileged commands (including shutdown, but not disc-
- connect), the normal level for another node.
- 8 - more privileged commands (including disconnect)
- 9 - local sysop privilege. DO NOT SET ANY REMOTE USER OR NODE TO THIS
- LEVEL.
+cd /spider
+cvs -z3 update -d
-If you are a sysop and you come in as a normal user on a remote connection -your privilege will automatically be set to 0. -
-
set/password <callsign> <string> Set a users password
-
-
The password for a user can only be set by a full sysop. The string -can contain any characters but any spaces are removed (you can type in -spaces - but they won't appear in the password). You can see the -result with STAT/USER. The password is the usual 30 character baycom -type password. -
-
set/sys_qra <locator> Set your cluster QRA locator
-
-
show/program Show the locations of all the included program modules
-
-
Show the name and location where every program module was load from. This -is useful for checking where you think you have loaded a .pm file from. -
-
shutdown Shutdown the cluster
-
-
Shutdown the cluster and disconnect all the users. If you have Spider -set to respawn in /etc/inittab it will of course restart. -
-
spoof <callsign> <command> Run commands as another user
-
-
This is a very simple yet powerful command for the sysop. It allows you to -issue commands as if you were a different user. This is very useful for the -kind of things that users seem to always get wrong.. like home_node for -example. -
-
stat/db <dbname> Show the status of a database
-
-
Show the internal status of a database descriptor. -
Depending on your privilege level you will see more or less information. -This command is unlikely to be of much use to anyone other than a sysop. -
-
stat/channel <callsign> Show the status of a channel on the cluster
-
-
Show the internal status of the channel object either for the channel that -you are on or else for the callsign that you asked for. -
Only the fields that are defined (in perl term) will be displayed. -
-
stat/msg <msgno> Show the status of a message
-
-
This command shows the internal status of a message and includes information -such as to whom it has been forwarded, its size, origin etc etc. -
-
stat/user <callsign> Show the full status of a user
-
-
Shows the full contents of a user record including all the secret flags -and stuff. -
Only the fields that are defined (in perl term) will be displayed. -
+
and your files will be updated. As above, remember to recompile the "C" client +if it has been updated (CVS will tell you) and restart if any of the perl scripts +have been altered or added, again, CVS will tell you.
+
You will find any changes documented in the /spider/Changes file.