Difference between revisions of "Talk:Nw65sp6"
(→XMGR.NLM memory leak)
|Line 98:||Line 98:|
== XMGR.NLM memory leak ==
== XMGR.NLM memory leak ==
XMGR.NLM seems to continually grab more memory. One
XMGR.NLM seems to continually grab more memory. One of my servers it is at 64825% its original size (367MB) and growing! All the servers, 13 total, are showing this, but the ones that have been up the longest the worst. The one I mention only has 3 users connecting to it and use it only minimally, if at all. This might be related to the following two issues.
== Unable to unlock screen saver ==
== Unable to unlock screen saver ==
Revision as of 16:46, 21 February 2007
Wohoo! Glad to see this updated so quicky! I hadn't gotten to it yet. Thanks! -- Gracal
You made it easy for me. I just copied the nw65sp5 page to start with. --Marcel 13:56, 17 November 2006 (MST)
- 1 CIFS Abend Issues Remain Unresolved, Post NW 6.5 SP6 CIFS Modules Installed
- 2 Broadcom issue with DELL 830, 2650, and 2850 and HP Proliant ML350R04 model 6/3400
- 3 Reproducable abend during Veritas backup
- 4 My problem with SP6.
- 5 phpmyadmin
- 6 BuildBanner() Failed
- 7 Slow access and NDPSGW problems.
- 8 Rsync Copy Error 6.5 SP2
- 9 Zen7 Imaging fail
- 10 XMGR.NLM memory leak
- 11 Unable to unlock screen saver
- 12 Cache memory allocator out of available memory
CIFS Abend Issues Remain Unresolved, Post NW 6.5 SP6 CIFS Modules Installed
I'm curious to know if anyone else has a NetWare CIFS in the PDC mode in production with stability issues - especially with ZenWorks 7 Middle Tier Server on NetWare?
We have no Windows or Samba servers in our Domain, only two NetWare 6.5 post SP6 servers with one configured to be the PDC and the other server joined to the PDC. Both servers will work fine for about 12 hours after configuration, then we see a cifsproxy.nlm Abend which kills CIFS completely on both servers.
Abend appears to be related to any random user login via a Windows 'Client-less' ZenWorks 7 Agent login via the NetWare Middle Tier server. Universal Password has been enabled for some users, and simple password only for others - same issue. We have also tried a similar configuration on both servers in Workgroup mode and we receive the same abend after a Middle Tier client login. Clients are a mix of fully patched, NCP 'Client-less' Windows 2000 and XP workstations, using the latest public ZfD agents and client utilities.
Novell Support (Netware & Zenwork Teams) are aware of this issue, but since they cannot consistently reproduce it there is no fix available (or forthcoming) at this time. We have decided to try OES SuSE Linux Samba instead, but as we are still learning Linux this is a very slow moving 'fix'.
Broadcom issue with DELL 830, 2650, and 2850 and HP Proliant ML350R04 model 6/3400
I had a weird issue that I have happen 3 times on 6 servers.
They all have the broadcom b57 driver. Everything was working fine and then sp6. On 3 of the 6 servers the driver board configuration, specifically the "board status" changed from "enabled" to "forced". "Line Speed: Automatic detection was still enabled. However, when I checked the Cisco switch, the speed did not match. I checked the switch and it was forced to 1000 FD. When I unplugged te ethernet cord and plugged it back in, it said that it was trying to link at 10 half duplex. I changed the board status from "Forced" to "enabled" rebooted, and everything was fine.
Also, still issues with iPrint even after all the beta patches.
Reproducable abend during Veritas backup
I'm testing SP6 on one of my internal servers before deployment to client sites.
This server is set up to backup across the wire to a remote server running Backup Exec.
After a few gigs on the backup, the SP6 patched server abends either in Apache or while trying to cleanup an address space. Since the only protected memory space on this server is Apache admin, it's again related to a problem with Apache.
I tried a few simple and obvious things, like updating the BIOS, switching from the CE1000.lan driver to the recommended E1000.lan driver, and installing the latest Backup Exec build. No joy.
Finally, I unloaded both occurrences of Apache and ran the backup. I got over 100GB backed up with no problems. Much joy.
So, I have a work around but I remain concerned about wider deployment.
My problem with SP6.
I have put this on two servers so far and have run into a significant delay if I select to update the iManager plugins. It will sit there for several hours before the updates are downloaded from the Novell site.
I see the same problem running the updates from iManager 2.5 or higher. It sees that the updates are there but certain ones take forever to download. I suspect that the Novell download site is overloaded or it is only accepting a limited number of connections. I wish someone at Novell would look into this problem because it isn't just me that has experienced this.
I've seen abends posted for phpmyadmin within the adminsrv namespace. I am experiencing this as well but so far the only suggestion has to been call Novell. Unfortunately we can't plunk down the $650 for the call. Here is the details if it helps:
1-02-2007 4:39:32 pm: SERVER-5.70-2187 [nmID=D0006] Removed address space because of memory protection violation Address Space: ADMINSRV Reason: Page Fault, Attempt to read from non-present page Running Thread: Apache_Worker 0 EIP: 0xF4995321 (PHP5LIB.NLM + 0xB7321) Access Locat
Gallwapa 15:33, 4 January 2007 (MST)
I am getting a "BuildBanner() failed" error on my logger screen after installing sp6. Any help would be appreciated.
Slow access and NDPSGW problems.
File access to server is slow and backups are also taking twice as long after applying SP6. NDPSGW.NLM is consuming memory until the server abends. We have applied the latest NDPSupdate from January 29 2007 (novpatch3761.zip) but no change.
Reinstalling SP6 solved the slow access to the server problem.
To solve the NDPSGW problem we created a new NDPS-manager object and created new ndps-printers and the ndpsgw.nlm started to consume memory when we created the npds-object of a sharp-printer. We located a sharp-ndps gateway via google http://www.sharp.marcocentral.net/USharpNDPSGateway031704.zip and installed it. It worked fine after we created a new sharp ndps-printer with the new gateway.
Rsync Copy Error 6.5 SP2
Rsync broken in Netware 6.5 SP 6? Go back to SP5.
I have two Netware 6.5 SP5 servers in my environment. One server Rsyncs to the other every night. I upgraded my backup server to Netware 6.5 SP6 to test SP6 before I deployed it to my main server. Everything seemed okay except Rsync broke. Rync appeared to copy the directories properly to the destination server. Rsync would start to copy the files. The usual temp file would be created and called for example, "file.txt.000224", then the file would be deleted and the next file would copy "file2.txt.000224", that file would be deleted and so on. At the end of the transfer when Rsync tried to rename each file to it's correct name (file.txt), it would report that the file did not exist, Error Code 23. If I checked the Salvage Files, all of the temp files were listed on the backup server directory. So it didn't seem to be a permissions issue. Additionally the primary server said the transfer was successful in it's log files. I checked my Rsync config files and they were the same as Gary Childer's in his "Rsync Cool Solutions" document. I am doing a simple Volume rsync without SSL. Then I hooked up a Windows workstation to Rsync as a test and got the same results. I tried replacing my Rsync*.nlm(s) in my \etc\rsync with the SP5 version on the other server. That did not help. I finally uninstalled SP6 on my backup server and the problem was fixed! Rsync is working again.
Zen7 Imaging fail
Error: PXE-T04 Illegal TFTP operation Error: PXE-E36 Error received from TFTP server
Server running 4 days after Update fine. Then TFTP stops with errors. Nameresolution on Netware Server fail? Only Solution: Reboot Server, then tftp working again some days.
XMGR.NLM memory leak
XMGR.NLM seems to continually grab more memory. One of my servers it is at 64825% of its original size (367MB) and growing! All the servers, 13 total, are showing this, but the ones that have been up the longest are the worst. The one I mention only has 3 users connecting to it and use it only minimally, if at all. This might be related to the following two issues.
Unable to unlock screen saver
We're unable to unlock the screen saver since upgrading to SP6. We can use the admin account and that works. I have discovered that if I change my password it works for about a day and then stops working.
Cache memory allocator out of available memory
I keep seeing error messages that Java was unable to request more memory. Server has 2GB memory. The java apps I have running are: APC Network shutdown, ZENLoader (ZfS agent component), and Dell openmanage. This is the error message:
2-21-2007 1:50:19 am: SERVER-5.70-0 [nmID=6001D] Cache memory allocator out of available memory.
2-21-2007 1:50:19 am: SERVER-5.70-0 [nmID=2000A] Short term memory allocator is out of memory. 28 attempts to get more memory failed. request size in bytes 135355754 from Module JAVA.NLM