I have a win2k server (fully patched) running 24x7 Scheduler 3.4.26.

I have a script looking for a semephor file once a minute.
Once it finds it, it runs the following job

//Publish site
SyncFTPDir( "Local", "www.webserver.com", "24x7 login", "24x7 pass", "C:\\www\\site", "//site", true, true, true, true )

Every now and then, the developers will rename a file or dir to a temp name (z_name) to replace it with an updated version for testing. If they forget to delete it before publishing, it goes to the live site. No big deal as nothing is linked to it. However, occasionally when they try to publish, 24x7 throws an error saying it can not delete z_name and the whole job fails. It seems that when it has an issue with the (now infamous in our office) Z files, it will have an issue with multiple Z files. It is randomly happening and can not be consistantly reproduced. I have checked the help file regarding Temporary Internet File Caching and have checked the settings listed there. One thing I have noted is that they all begin with Z in some form (z_filename, zz_filename, etc). I don't know if the name is part of the cause, or if it is just coincidence and it is the actions or the developers (nameing/renaming/deleting), but as some of the files that are causing problems haven't been touched in days/weeks, timestamp or filecaching issues seem unlikely. Permissions also seem to be ruled out as it works most of the time.

I have had this same problem in the past and have upgraded 24x7 as well as changed hardware. It should also be noted that the script is publishing from one subnet through a firewall to another subnet. The firewall logs show no unusual activity or alerts.

I am also having issues with 24x7 stop looking for semephor files (though 24x7 is still running with no errors). This is another recurring issue that I will post separately.

Have you checked ftp.log file? Is it trying to delete Z_something from the web server which isn't there or it's there and can't be deleted?

: I have a win2k server (fully patched) running 24x7 Scheduler 3.4.26.

: I have a script looking for a semephor file once a minute.
: Once it finds it, it runs the following job

: //Publish site
: SyncFTPDir( "Local", "www.webserver.com", "24x7
: login", "24x7 pass", "C:\\www\\site",
: "//site", true, true, true, true )

: Every now and then, the developers will rename a file or dir to a temp name
: (z_name) to replace it with an updated version for testing. If they forget
: to delete it before publishing, it goes to the live site. No big deal as
: nothing is linked to it. However, occasionally when they try to publish,
: 24x7 throws an error saying it can not delete z_name and the whole job
: fails. It seems that when it has an issue with the (now infamous in our
: office) Z files, it will have an issue with multiple Z files. It is
: randomly happening and can not be consistantly reproduced. I have checked
: the help file regarding Temporary Internet File Caching and have checked
: the settings listed there. One thing I have noted is that they all begin
: with Z in some form (z_filename, zz_filename, etc). I don't know if the
: name is part of the cause, or if it is just coincidence and it is the
: actions or the developers (nameing/renaming/deleting), but as some of the
: files that are causing problems haven't been touched in days/weeks,
: timestamp or filecaching issues seem unlikely. Permissions also seem to be
: ruled out as it works most of the time.

: I have had this same problem in the past and have upgraded 24x7 as well as
: changed hardware. It should also be noted that the script is publishing
: from one subnet through a firewall to another subnet. The firewall logs
: show no unusual activity or alerts.

: I am also having issues with 24x7 stop looking for semephor files (though
: 24x7 is still running with no errors). This is another recurring issue
: that I will post separately.

: Have you checked ftp.log file? Is it trying to delete Z_something from the
: web server which isn't there or it's there and can't be deleted?

Here is the FTP log. I've replaced the USER with 24x7login and the IP with 1.1.1.1

08:32:56 1.1.1.1 [8833]PASS - - 230
08:32:56 1.1.1.1 [8834]USER 24x7login - 331
08:32:56 1.1.1.1 [8834]PASS - - 230
08:32:56 1.1.1.1 [8834]DELE z_application.cfm - 250
08:32:56 1.1.1.1 [8834]DELE z_index.cfm - 250
08:32:56 1.1.1.1 [8834]DELE zz_index.cfm - 250
08:32:56 1.1.1.1 [8834]DELE zzz_index.cfm - 250
08:32:56 1.1.1.1 [8835]USER 24x7login - 331
08:32:56 1.1.1.1 [8835]PASS - - 230
08:37:08 1.1.1.1 [8836]USER 24x7login - 331
08:37:08 1.1.1.1 [8836]PASS - - 230
08:37:08 1.1.1.1 [8837]USER 24x7login - 331

Here is the 24x7 log. I've replace machine name with xxxxxx

11/18/2005 08:48:30:312 xxxxxx 105 Pub Site Job finished.
11/18/2005 08:45:47:906 xxxxxx 105 Pub Site Job started.
11/18/2005 08:44:16:109 xxxxxx 105 Pub Site An error occurred while executing 24x7 script: Line 2: Error deleting remote subdirectory or files in subdirectory
11/18/2005 08:42:00:687 xxxxxx 105 Pub Site Job started.
11/18/2005 08:37:52:593 xxxxxx 105 Pub Site An error occurred while executing 24x7 script: Line 2: Error deleting remote subdirectory or files in subdirectory
11/18/2005 08:36:42:437 xxxxxx 105 Pub Site Job started.
11/18/2005 08:35:25:14 xxxxxx 105 Pub Site An error occurred while executing 24x7 script: Line 2: Error deleting remote subdirectory or files in subdirectory
11/18/2005 08:35:15:984 xxxxxx 105 Pub Site Job started.
11/18/2005 08:34:24 xxxxxx -- 24x7 Scheduler 24x7 Scheduler starting...
11/18/2005 08:34:19 xxxxxx -- 24x7 Scheduler [24x7 service] 24x7 Scheduler terminating...
11/18/2005 08:32:53:25 xxxxxx 105 Pub Site Job has been disabled!
11/18/2005 08:32:50:39 xxxxxx 105 Pub Site Semaphore file(s) C:\www\pubflags\pub_flag.txt found. Submitting job to the queue.

Are these valid files? I mean do get references files like zz_index.cfm are actually on the web sever side before the job starts? If yes, are they deleted during job run? If yes, again, what's left? Which file exactly is causing the problem in the provided log?

: Here is the FTP log. I've replaced the USER with 24x7login and the IP with
: 1.1.1.1

: 08:32:56 1.1.1.1 [8833]PASS - - 230
: 08:32:56 1.1.1.1 [8834]USER 24x7login - 331
: 08:32:56 1.1.1.1 [8834]PASS - - 230
: 08:32:56 1.1.1.1 [8834]DELE z_application.cfm - 250
: 08:32:56 1.1.1.1 [8834]DELE z_index.cfm - 250
: 08:32:56 1.1.1.1 [8834]DELE zz_index.cfm - 250
: 08:32:56 1.1.1.1 [8834]DELE zzz_index.cfm - 250
: 08:32:56 1.1.1.1 [8835]USER 24x7login - 331
: 08:32:56 1.1.1.1 [8835]PASS - - 230
: 08:37:08 1.1.1.1 [8836]USER 24x7login - 331
: 08:37:08 1.1.1.1 [8836]PASS - - 230
: 08:37:08 1.1.1.1 [8837]USER 24x7login - 331

: Here is the 24x7 log. I've replace machine name with xxxxxx

: 11/18/2005 08:48:30:312 xxxxxx 105 Pub Site Job finished.
: 11/18/2005 08:45:47:906 xxxxxx 105 Pub Site Job started.
: 11/18/2005 08:44:16:109 xxxxxx 105 Pub Site An error occurred while executing
: 24x7 script: Line 2: Error deleting remote subdirectory or files in
: subdirectory
: 11/18/2005 08:42:00:687 xxxxxx 105 Pub Site Job started.
: 11/18/2005 08:37:52:593 xxxxxx 105 Pub Site An error occurred while executing
: 24x7 script: Line 2: Error deleting remote subdirectory or files in
: subdirectory
: 11/18/2005 08:36:42:437 xxxxxx 105 Pub Site Job started.
: 11/18/2005 08:35:25:14 xxxxxx 105 Pub Site An error occurred while executing
: 24x7 script: Line 2: Error deleting remote subdirectory or files in
: subdirectory
: 11/18/2005 08:35:15:984 xxxxxx 105 Pub Site Job started.
: 11/18/2005 08:34:24 xxxxxx -- 24x7 Scheduler 24x7 Scheduler starting...
: 11/18/2005 08:34:19 xxxxxx -- 24x7 Scheduler [24x7 service] 24x7 Scheduler
: terminating...
: 11/18/2005 08:32:53:25 xxxxxx 105 Pub Site Job has been disabled!
: 11/18/2005 08:32:50:39 xxxxxx 105 Pub Site Semaphore file(s)
: C:\www\pubflags\pub_flag.txt found. Submitting job to the queue.

: Are these valid files? I mean do get references files like zz_index.cfm are
: actually on the web sever side before the job starts? If yes, are they
: deleted during job run? If yes, again, what's left? Which file exactly is
: causing the problem in the provided log?

In this instance, the files z_index.cfm, zz_index.cfm, and zzz_index.cfm were on the webserver. When 24x7 tried to delete them (they were removed from the source), the job failed. When I ran in it debug mode with trace on, I saw the message "could not delete file z_index.cfm. Access denied". Once I removed z_index.cfm from the target server, and tried again, it gave the same results except it had issues with zz_index.cfm. This continued with all the z_ files in that directory and one other directory. It got to the point where I did a search for all files/direcotries that began with z and removed the z_ files. It then ran fine.

I was able to log into the ftp server using 24x7's login/pass and delete the files using IE, but the program wouldn't do it.

It seems strange that sometimes it has no issues at all with z_ files and sometimes it does. But when it has issues with one z_ file, it doesn't like dealing with any z_ files.

: In this instance, the files z_index.cfm, zz_index.cfm, and zzz_index.cfm were
: on the webserver. When 24x7 tried to delete them (they were removed from
: the source), the job failed. When I ran in it debug mode with trace on, I
: saw the message "could not delete file z_index.cfm. Access
: denied". Once I removed z_index.cfm from the target server, and tried
: again, it gave the same results except it had issues with zz_index.cfm.
: This continued with all the z_ files in that directory and one other
: directory. It got to the point where I did a search for all
: files/direcotries that began with z and removed the z_ files. It then ran
: fine.

: I was able to log into the ftp server using 24x7's login/pass and delete the
: files using IE, but the program wouldn't do it.

: It seems strange that sometimes it has no issues at all with z_ files and
: sometimes it does. But when it has issues with one z_ file, it doesn't
: like dealing with any z_ files.

It has happed again, but this time with a regularly named file (ie not a z_ file).
The symptoms are similar. The files were renamed or deleted multiple times on the source and then the job ran once, threw an error (see below), and was disabled.

Error deleting remote subdirectory or files in subdirectory " /FCKeditor". FTP Error. File search operation failed: Error code 18

The last time the job ran was the previous day, when the FCKeditor directory was first created on both the source and destination servers.

The FCKeditor directory was then moved into another directory (one level deeper in the file structure) on the source

The job ran and choked on the original FCKeditor on the dest server.

So it somehow gets directory listing from the cache, in other words on recursive runs uses cached data rather then current data.

This means that you haven't turned off file caching option in your Windows settings. See Control Panel -> Internet Settings ->Temporary Files -> Settings -> "Check For Newer Versions of stored pages" must be set to "Every visit to the page."

Also, I suggest you uncheck "disable job on error" option in that job's properties. So if it fails you don't need to re-enable it.

: It has happed again, but this time with a regularly named file (ie not a z_
: file).
: The symptoms are similar. The files were renamed or deleted multiple times on
: the source and then the job ran once, threw an error (see below), and was
: disabled.

: Error deleting remote subdirectory or files in subdirectory "
: /FCKeditor". FTP Error. File search operation failed: Error code 18

: The last time the job ran was the previous day, when the FCKeditor directory
: was first created on both the source and destination servers.

: The FCKeditor directory was then moved into another directory (one level
: deeper in the file structure) on the source

: The job ran and choked on the original FCKeditor on the dest server.

: So it somehow gets directory listing from the cache, in other words on
: recursive runs uses cached data rather then current data.

: This means that you haven't turned off file caching option in your Windows
: settings. See Control Panel -> Internet Settings ->Temporary Files
: -> Settings -> "Check For Newer Versions of stored pages"
: must be set to "Every visit to the page."

: Also, I suggest you uncheck "disable job on error" option in that
: job's properties. So if it fails you don't need to re-enable it.

When we first started this thread, I had double checked that. My caching is turned off per instructions. That's why I am so puzzled by this. In addition, I'm not sure why all the other jobs stop looking for the semaphore files when one fails, as it seems to be happening. I will continue to monitor this and see of I can reproduce the "All jobs stop" thing.

I will uncheck the disable of fail though and see what happens.

Do your FTP server or proxy/ISA servers cache any results? What kind of FTP server do you run?

: When we first started this thread, I had double checked that. My caching is
: turned off per instructions. That's why I am so puzzled by this. In
: addition, I'm not sure why all the other jobs stop looking for the
: semaphore files when one fails, as it seems to be happening. I will
: continue to monitor this and see of I can reproduce the "All jobs
: stop" thing.

: I will uncheck the disable of fail though and see what happens.

: Do your FTP server or proxy/ISA servers cache any results? What kind of FTP
: server do you run?

We use IIS 5's built in FTP server on all of our FTP servers that 24x7 sync with. I have looked at the config, but there does not seem to be a way to see whether or not it caches results like directory listing. They are using MSDOS style directory listing as opposed to Unix style.

Take a look at MS KB Article Q182626

http://support.microsoft.com/kb/q182626/

It explains file caching features in IIS and describes how to turn it off. Hope this help.

: We use IIS 5's built in FTP server on all of our FTP servers that 24x7 sync
: with. I have looked at the config, but there does not seem to be a way to
: see whether or not it caches results like directory listing. They are
: using MSDOS style directory listing as opposed to Unix style.

: Take a look at MS KB Article Q182626

: http://support.microsoft.com/kb/q182626 /

: It explains file caching features in IIS and describes how to turn it off.
: Hope this help.

It refers to IIS4, but I'll try it and let you know how it works out.

: It refers to IIS4, but I'll try it and let you know how it works out.

I have made the changes, but still no luck.
The source and dest servers have a dir named Pages, which has not been touched for months. Today, a developer have removed that dir and added/modified a few files on the source. We published (just once) and it threw an error:

Error deleting remote subdirectory or files in subdirectory "//server/path/Pages". Ftp Error. File Search operation failed: The requested operation cannot be made on the FTP session handle because the operation is already in progress.

The file is still on the dest server. I can delete it just fine using IE (via the FTP service).

Weird. I am looking what else could be causing this caching. I will get back to you later today

: I have made the changes, but still no luck.
: The source and dest servers have a dir named Pages, which has not been
: touched for months. Today, a developer have removed that dir and
: added/modified a few files on the source. We published (just once) and it
: threw an error: Error deleting remote subdirectory or files in
: subdirectory "//server/path/Pages". Ftp Error. File Search
: operation failed: The requested operation cannot be made on the FTP
: session handle because the operation is already in progress.

: The file is still on the dest server. I can delete it just fine using IE (via
: the FTP service).

Here is what you can try as an alternative method.

In the Windows explore map your FTP server connection as network drive (Tools/Map Network Drive menu; then in the share name type ftp://[whatever]; click "Connect using" link and enter FTP user id and password; save the mapping
In 24x7 job replace SyncFTPDir with SyncLocalDir and for the remote files instead of FTP use mapping created in step 1.

Please let me know if you need help with that. Note that SyncLocalDir does not need server-user-password parameters.

: Weird. I am looking what else could be causing this caching. I will get back
: to you later today