ticket #2425

[Nuxwin]

ispcp-apache logger was rewritten. i'll commit later. The problem will be fixed.

About the problem (subject of this thread), if you have not the same problem with the update script, it's not a sleep statement that must be added. With this last information, we can now inspect correctly (lock issue).

Note: The limit 1024 can be changed.

And please Daniel you know me so, calm down. I don't understand why your are nervous. I've proposed you IRC session but you didn't answer.

---

No pm activated so:

sci2tech by PM Wrote:Never received any invitation to discuss anything from nobody. And I`m not nervous, my english is even worse because I stopped using since 1 year, and this maybe make you / anybody else believe that I`m nervous . But I dislike to see that every action I take is seen as provocation or some flame try. Was just a simple error report, and as soon a I opened that ticket my trac account stopped to work, that’s why I used forum. My mistake it seems. Anyway just for the record that limit can be changed with ulimit (kind of tricky) or recompiling apache. I did used both, and you can see this by a simple forum search

Read http://isp-control.net/forum/thread-1142...l#pid86386

---> if you want, maybe we can schedule an IRC session to talk about this.

Your action are not taken like provocations. For the trac issue, I've not removed your account. You should ask Malte for this issue. It's a strange story. You know, I have nothing against you. I would have preferred to continue working with you but unfortunately, you chose to leave the team. I resent you a bit for that.

[sci2tech]

I did not say that anybody did something with my trac account. It just stopped to work (send Location: http://www.isp-control.net/ispcp/prefs/account
in infinite loop). PM is enabled now. I`ll send you pm with my account data to check because it seems not to believe). Anyway I could replicate with a load of 0.25 and i could not with with a load of 23. And this was not happened with older version even if server was more loaded. You decide what`s next. I`ll just let it go.

[Nuxwin]

Please Daniel

Marc and me are currently inline to do some stress test on ispCP. What the FS used on your server ? ext3, another ? You use NFS or not ?

[sci2tech]

mount
/dev/hda1 on / type ext3 (rw,errors=remount-ro)

Linux 2.6.26-2-686 #1 SMP Tue Mar 9 17:35:51 UTC 2010 i686 GNU/Linux

cat /etc/issue
Debian GNU/Linux 5.0
If more needed just say.

[sci2tech]

Found one serious issue in ispcp-rqst-mngr that can explain this behavior. Domains are deleted after servers are restarted (ispcp-dmn-mngr is runing after ispcp-serv-mngr, config files are altered, and since apache2ctl graceful way of treating this signal (USR1) it is possible to catch ispcp.conf in plain process of copying. Guess you do not need a patch for this, so from my point of view problem is solved for now. That delay gives a chance to finish the script, otherwise was not useful.

[Nuxwin]

ah great
thank. I'll take care of that

Edit:

Same here -> http://isp-control.net/forum/thread-11459.html ???

[sci2tech]

As posted by author, is not.
Patch:

Code:

Index: ispcp-rqst-mngr
===================================================================
--- ispcp-rqst-mngr    (revision 3238)
+++ ispcp-rqst-mngr    (working copy)
@@ -458,42 +458,7 @@
        }
    }

-    if ($dmn_count > 0 || $sub_count > 0 || $als_count > 0 || $mail_count > 0 ||
-        $alssub_count > 0) {

-        $rs = sys_command(
-            "$main::ispcp_serv_mngr $dmn_count $sub_count $als_count " .
-            "$mail_count $alssub_count 1>$main::ispcp_serv_mngr_stdout " .
-            "2>$main::ispcp_serv_mngr_stderr"
-        );
-
-        if ($rs != 0) {
-            if (-e $main::ispcp_serv_mngr_el) {
-                my $rs1;
-
-                # removing 'sys_command()' error message;
-                pop_el(\@main::el);
-
-                ($rs1, $rdata) = get_el_error($main::ispcp_serv_mngr_el);
-                return $rs1 if ($rs1 != 0);
-
-                push_el(\@main::el, $main::ispcp_serv_mngr, $rdata);
-            }
-
-            if (!defined($main::engine_debug)) {
-                del_file($main::ispcp_serv_mngr_stdout);
-                del_file($main::ispcp_serv_mngr_stderr);
-            }
-
-            return $rs;
-        }
-
-        if (!defined($main::engine_debug)) {
-            del_file($main::ispcp_serv_mngr_stdout);
-            del_file($main::ispcp_serv_mngr_stderr);
-        }
-    }
-
    #
    # .htusers management BEGIN.
    #
@@ -856,6 +821,42 @@
    # Domains data deletion END
    #

+    if ($dmn_count > 0 || $sub_count > 0 || $als_count > 0 || $mail_count > 0 ||
+        $alssub_count > 0) {
+
+        $rs = sys_command(
+            "$main::ispcp_serv_mngr $dmn_count $sub_count $als_count " .
+            "$mail_count $alssub_count 1>$main::ispcp_serv_mngr_stdout " .
+            "2>$main::ispcp_serv_mngr_stderr"
+        );
+
+        if ($rs != 0) {
+            if (-e $main::ispcp_serv_mngr_el) {
+                my $rs1;
+
+                # removing 'sys_command()' error message;
+                pop_el(\@main::el);
+
+                ($rs1, $rdata) = get_el_error($main::ispcp_serv_mngr_el);
+                return $rs1 if ($rs1 != 0);
+
+                push_el(\@main::el, $main::ispcp_serv_mngr, $rdata);
+            }
+
+            if (!defined($main::engine_debug)) {
+                del_file($main::ispcp_serv_mngr_stdout);
+                del_file($main::ispcp_serv_mngr_stderr);
+            }
+
+            return $rs;
+        }
+
+        if (!defined($main::engine_debug)) {
+            del_file($main::ispcp_serv_mngr_stdout);
+            del_file($main::ispcp_serv_mngr_stderr);
+        }
+    }
+
    push_el(\@main::el, 'mngr_engine()', 'Ending...');

    0;