2
0
mirror of https://github.com/xcat2/xcat-core.git synced 2025-06-02 19:40:10 +00:00

273 Commits

Author SHA1 Message Date
huweihua
e6c157b95b fix bug#155 There are errors when restart xcatd for UDP service's bug 2015-10-15 03:45:30 -04:00
huweihua
58e825691f fix bug#203 xdsh could not recognize single quote mark 2015-09-27 23:10:11 -04:00
huweihua
1ccf994d37 modify some code depending on xiaopeng's comments 2015-09-21 14:21:33 +08:00
huweihua
c7e430ae34 fix bug#160-165 and make commands.log support rotate mechanism 2015-09-21 14:21:33 +08:00
huweihua
fc53bb3b3c fix bug#140 offline trouble shooting time stamp wrong 2015-09-01 22:38:56 -04:00
huweihua
e7c5580a10 change fork ot xfork for command log process 2015-08-28 04:12:14 -04:00
huweihua
6414da0d34 modify indent 2015-08-28 02:55:49 -04:00
huweihua
efa45a81d9 command log process support restartxcatd and fix some response parsing error 2015-08-28 02:20:08 -04:00
huweihua
4058f6b60a fix typo, lost a { 2015-08-23 22:37:51 -04:00
huweihua
69ebde1728 fix some quit problem of command log writer process 2015-08-20 20:53:57 -04:00
huweihua
efb4aff9a5 enhance command log process for handling TERM signal 2015-08-18 23:45:36 -04:00
huweihua
4dcbe0ea48 modify many things depending on xiao peng and victor's comments 2015-08-18 04:17:37 -04:00
huweihua
8cd8818d5b record command output and add trace information 2015-08-13 23:27:02 -04:00
WangXiaoPeng
921b7d5ea3 defect 4759: Fix the issue that ssl_listener used up all cpu. 2015-07-24 01:17:36 -04:00
Jarrod Johnson
f00a23aa58 Ignore a dead UDP handler
If the UDP handler is killed (for some reason),
allow the SSL listener to continue.
2015-07-23 14:19:05 -04:00
GONG Jie
ce9430d280 Fix sourceforge bug #4691 Use a consistent log tag for xCAT log 2015-07-02 14:58:53 +08:00
WangXiaoPeng
0827547cea fixed the issue that token expire time was one month ealier than the correct time 2015-06-18 04:15:21 -04:00
ertaozh
7a1fb1afc3 fix hardware discovery issue that the findme request can be larger than 1500 bytes 2015-06-11 03:19:08 -04:00
Jarrod Johnson
6916b37234 Reduce communication between SSL and UDP
Previously, every TLS connection caused a communication with UDP
and this was bad.  Then every UDP connection caused a communication
and this was better.  This takes things a little further by only
doing the communication if there is a known interested party now.
2015-05-27 10:39:52 -04:00
Jarrod Johnson
fdc408f270 Decrease forecast as presumed real clients connect
The fudge factor is now decremented according to real number of clients counted
toward sslclient count.  Also, have the overload case break into flow control
communication as well (though it wouldn't have a good answer for an interested client)
2015-05-27 10:35:10 -04:00
Jarrod Johnson
49c809e031 Reorganize UDP <-> SSL rework
Because SSL actually was relying upon accept() to hold indefininetly,
it would hang up the UDP worker indefinitely, blocking requests.
Fix this by driving the loop iteration through select() rather than
accept() and iterating through each file handle only when select()
says they are ready.
2015-05-27 10:23:37 -04:00
Jarrod Johnson
5e248b0b0d Rework traffic control to not involve UDP every SSL connection
If the UDP monitor was busy, the SSL connections could back up.
Rework it so that the UDP monitor would get hung up by the SSL
process instead, if *any* of them get backed up.  However, the
nature of the loop in the SSL client is sufficiently lightweight
that it shouldn't hang up the UDP process under any reasonable
circumstance.  If it did hang it up though, it'll be less hung up
than discovery does already.
2015-05-26 17:03:19 -04:00
WangXiaoPeng
71a764ed96 defect 4644: change in xcatd and Client.pm to fix two issues: 1. if the plugin process run certain command like 'omshell' dhcp.pm and which hang, this will cause the plugin process cannot be finished with ctrl^c; 2. the ctrl^c cannot be passed to service node. 2015-04-21 05:15:43 -04:00
mellor
85f758bd0d defect 4587 - fix xcatd msg for undefined variable: Use of uninitialized value $line in length 2015-03-09 13:56:47 -04:00
ertaozh
4dd21f1775 fix bug 4555: lsslp -s PBMC does not honor -n flag 2015-02-05 04:35:16 -05:00
mellor
874f77afed defect 4082 - change previous fix to disable warnings when calling setlogsock to work with older versions of perl too 2015-02-04 08:00:03 -05:00
mellor
b34537ec2e defect 4529: send msg to install monitor process to reload perl modules when rescanplugins called 2015-01-21 16:08:34 -05:00
daniceexi
ad68658c3c defect 4500: skip the PIPE signal when call xcatd->relay_fds() to receive messages from plugin processes 2014-12-29 02:47:43 -05:00
daniceexi
476f9f7b94 defect 4442: fix the issue in xcatd that new UDP won't exit with error message when it cannot get socket during the xcatd fast restart 2014-11-28 06:57:36 -05:00
daniceexi
900963b1db defect 4375: fix the xcatd which has issue to receive request from xcatd which is longer than about 160,000 bytes 2014-11-20 07:08:36 -05:00
daniceexi
3a4cefe53b defect 4336,4352: in xcatd, for aix, make all the process (fix for udp process) to be stopped when received TERM singal from stopsrc command to the main process 2014-11-19 07:44:49 -05:00
daniceexi
df970f4cb0 defect 4384: fix the issue that copycds command cannot be ctrl+c cleanly. The root cause was INT signal was not received by the plugin process after it sends out from relay_fds. The fix is to send both INT and TERM. TERM can be received in this case. 2014-11-13 03:24:33 -05:00
daniceexi
f0b9cac61c defect 4304: fixed several things: 1. how to handle ctrl+c; 2. handle the plugin process cannot be stopped by TERM signal to process group SSL_Listener. (make the plugin process locate in the SSL_listener process group); 3. Make DB Access process exits later than in processing request. 2014-11-06 06:24:50 -05:00
mellor
7e8c02bf07 defect 4304 - fixed some more interprocess communication problems 2014-10-22 17:20:07 -04:00
mellor
2511cebef8 defect 4284 - fixed xcatd attempt to write to broken pipe which caused DB processes to not get killed 2014-10-14 20:48:51 -04:00
zhaoertao
aea40b0226 fix an issue about hardware discovery process: the sub process can not receive findme request immediately 2014-10-04 23:20:48 -07:00
mellor
4e008d7895 enhance for p8 hardware discovery 2014-09-26 03:54:52 -07:00
mellor
cea7f53e91 defect 4229 add new rescanplugins support to xcatd 2014-07-29 11:19:55 -04:00
mellor
d6a55aa3dd defect 4082 - disable warnings when calling setlogsock 2014-06-11 13:20:42 -04:00
daniceexi
6bb22d2342 added the token authentication mechanism in the xcatd 2014-04-07 07:57:57 -04:00
mellor
a698e1757d defect 4020 - correctly handle results from gethostbyaddr in xcatd installmonitor 2014-03-31 17:50:05 -04:00
daniceexi
adf723db25 fix the issue that xcatd hangs when reading the request which is formatted in one line like litefile request from statelte CN 2014-01-08 07:42:35 -05:00
daniceexi
e7dc46a03a defect 3909: make xcatd loads xCAT::Enabletrace by require instead of use to save resource 2014-01-02 02:05:37 -05:00
zhaoertao
8bc0488764 fix bug 3889 xcatd not running preprocess for multiple plugins when mgt=ipmi 2013-11-21 00:52:08 -08:00
daniceexi
399e71a4c0 defect 3789: fixed the issue that xcatdebug command kill the xcatd ssl listening process. The fix is that just remove the handling debug message from plugin modules since xcatd loads plugin modules for each command run, don't need xcatdebug for plugin any more. 2013-10-24 10:41:48 -04:00
Jarrod Johnson
857af4d498 Send a positive acknowledgement of flow request 2013-10-23 11:15:04 -04:00
lissav
b9144fc028 comment out for now disable buffering on the socket 2013-10-15 14:24:43 -04:00
Jarrod Johnson
62b436be08 Disable storable encoding for client
Clever things can live in Storable payloads that can essentially be remote execution.
Disable that for now.
2013-10-11 15:57:29 -04:00
jbjohnso
40a65cc564 Fix problem where xcatd would error on end of stream
git-svn-id: https://svn.code.sf.net/p/xcat/code/xcat-core/trunk@17022 8638fb3e-16cb-4fca-ae20-7b5d299a9bcd
2013-07-25 13:27:37 +00:00
jbjohnso
e5dd20f15e Clear $1 prior to calling plugins
xcatd now might let $1 leak into the plugins.  Ideally, the downstream code
would check match before relying upon $1 to be undef, but better to
avoid unintuitive behavior.

git-svn-id: https://svn.code.sf.net/p/xcat/code/xcat-core/trunk@16997 8638fb3e-16cb-4fca-ae20-7b5d299a9bcd
2013-07-17 18:59:10 +00:00