[color purple]NWagent causes ARCser 1

[LolaRed]

NWagent causes ARCserve to freeze since applying cumulative SP2 to ARCserve 7 and to the NWagent on Netware 5.11 SP5 servers. As soon as the ARCserve backup job accesses the server running NWagent, ARCserve simply stops and the backup job never finishes.

Has anyone else experienced this particular issue?

CA tech support wants me to troubleshoot their patch bugs for them. I know what caused it - the cumulative patch. I just want them to fix it.

Also, when I backup my GroupWise server, not using the GroupWise agent and also not using NWagent to pre/post gstop/gstart before & after the backup job (re: the first issue), it appears to be getting all of the post office files - not skipping any. Why do I even need any agent in the first place? Am I missing something here?

 

[LoYBoY]

Hi LolaRed,

I think I am having roughly the same problem. I am running the same config as you (NW5.1 Spv5 servers, Arcserve v7 fully patched).

I will try and be specific as I can here to describe my problem:

From a clean arcserve database:

Create a GFS job, Full back up on a Friday, no action on S,S incremental job on M,T,W,T

The full backup on a Friday works (when I say works it backs up data from the selected nodes in the job but sometimes not everything on the server!)

Incremental Job on a Monday starts - formats the blank tape correctly - Arcserve contacts the first node in the job - this node receives the job request from arcserve (this can be viewed from the Client Agent for Netware screen - Real Time Statistics on a RConsole to server) - THEN NOTHING!!

Arcserve scheduler screen displays job as active. I have tried cancelling this node (using RConsole, arcserve scheduler screen and moving on to the next node within the Incremental job but the same symptoms persist.

Cancelled Mon incremental job, same issue again on Tuesday incremental job.

I had started to look at the Arcserve database / Btrieve settings until I read your post, at least this gives a insight to what may be causing the problem. I don't fancy removing the cumulative patches (and all the others related!).

Any ideas??

 

[TheLad]

Have you got the TSA agents loaded (TSA500 and TSANDS)? Are you backing up via IP or IPX? Do you have any NDS replicas on the server running Arcserve?

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[LolaRed]

Hi LoYBoY. Well, it's nice to know I'm not alone with this issue. I don't fool around with incrementals anymore. ARCserve doesn't stick to that for very long - plays nicely for a few days, then falls back to full dumps before it's supposed to. I just do a full dump every night - what the heck, as long as it's done before morning.... My current CA tech support has tried to keep me happy by requesting logs, .ncfs, etc. and keeps reassuring me (with the exact same form e-mail- imagine that!) that he's studying all my information. I've had issues with ARCserve in the past and usually get my answers right here from these fine tek-tip cruisers.

Here's what I've done so far.
1. I found and implemented good btrieve info at this CA URL:

http://support.ca.com/techbases/as61/10133.html

2. Since I'm using a DLT autochanger Tape Library, I removed NWTAPE.cdm and another novell tape driver (I'm at home right now...) from Startup.NCF and renamed the files. This was suggested by CA tech support because they've recently perfected their own drivers but there's a conflict if Novell drivers are loaded too.

3. Hi TheLad. These next items are in response to your excellent questions. I have tried backing up with both IP and IPX. Same problem. TSA500 and TSANDS are both loaded on all servers to be backed up. And yes, there are Read/Write replicas of the other servers on the ARCserve server. I assume they don't have to be Masters.... but thanks for asking, TheLad. I appreciate the tips.

4. My most recent suggestion by CA tech support is what I said before. Remove the cumlative SP2, put on QO05996 (SP1) and a couple of others (again, I'm at home - don't remember which ones). The point is, I already know that will work, my last ARCserve setup was before SP2 and I didn't have this problem (same hardware) - but of course there were other issues.... Sigh, 2 giant steps back....

5. Currently, I'm not using NWagent (to do a pre/post gstop & gstart before & after backing up the mail server) on my GroupWise server. My backups seem fine (as long as I keep agents out of the mix) and the post office seems to be fully backed up. What am I missing here? I thought that the post office could not be properly backed up without either the GroupWise agent or Netware agent to locally shut down the post office first. I've been able to restore NWguard, ofuser files, & other random post office files. It doesn't show any skipped files in the log.

6. Here's why I'm so annoyed by this problem. When ARCserve hits the NWagent on the mailserver, I have the agent perform a local PRE-backup job command, GSTOP, to shut down the GroupWise agents. Then the job hangs all night and GroupWise is down (GSTART never happens). Some of my users begin their workday 2 hours before I do. So this has caused a decrease in user confidence in the mail system - when it first began. It's pretty bad when you're trying to encourage them to use the GroupWare tools and they wonder why they should bother when they don't feel confident that it will be available. I'm not interested in testing this irritating issue on my live system. AND, my weekends have been wasted because short tests work fine (go figure!).

Ok, my rant is over. I will keep you posted when I've got something worthwile to share.

 

[8arcserve]

"NWagent causes ARCserve to freeze since applying cumulative SP2 to ARCserve 7 and to the NWagent on Netware 5.11 SP5 servers"
I have encountered this error from day one, after updating Arcserve 7 Netware using patchs Q33421 and Q33425.

And this fault seems to be intermitient, some nights it works. Other nights it just hangs on jobs and other times comes up with
E3149/E3148 errors (failed to begin new session in arcserve database). The last errors are btrieve errors. See link:

http://support.ca.com/techbases/as61/10133.html

I have found a work around for all tht above errors, simply use the old NWAGENT QO28844. Which seems to work fine. No btrieve errors or hangs.

Also which version of groupwise are you using, as I backup a groupwise 5.5 server using both the gw open file agent and nwagent (you need both if backing up using IP only) and have had nothing but trouble since updating arcserve 7. Which is strange as arcserve 6.6 work like a dream..?

 

[LoYBoY]

LolaRed - any chance you could expand on no.2 from your post above?

I also use a DLT autochanger. NWTAPE.CDM details as follows:

Novell 5.00 generic tape CDM
Version 1.00m 2 November 1999.

Do you have something different??

Thanks.

loyboy55@hotmail.com

8arcserve - I will try rolling back NWagent on one of my nodes to test - thanks.

 

[8arcserve]

LoYBoY
For info on NWTAPE.CDM read the following links.

http://support.novell.com/cgi-bin/search/searchtid.cgi?/10070665.htm

http://support.ca.com/techbases/as61/18032.html

 

[LoYBoY]

8Arcserve - I take it your arcserve server is patched with Q033421 and its only the NWagent patch you rolled back to QO28844?

The reason I ask is as follows:

NOTE: Please be aware that this patch contains
NetWare Client Agent updates that are only
compatible with the latest cumulative patch
(QO33421.caz) for the ARCserve 7 for NetWare Base
product available on the ARCserve 7.0 for NetWare
Patches and Updates web page. As such, a machine
that has been updated with these Client Agent files
should not be used in conjunction with an ARCserve
host server running on a different NetWare server
that has not had QO33421 applied.

Do you see what I am getting at?

Thanks.

Lb.

 

[8arcserve]

Yes I'm aware of the conflicts with running patched and unpatched versions of arcserve.
I originally patched all the netware servers running arcserve 7 host with QO33421 and also patched all the servers running the NWAGENT.NLM with QO33425.
After which about half the servers running the QO33425 patched NWAGENT.NLM started to hang jobs on an intermitient basis.
I rolled back to the QO28844 patch on the problem servers and have had no problems with them since. I know this is not what CA intended but its the only way i've got the agents to work on certain servers.

 

[LolaRed]

Hi LoYBoY. To expand on #2 (as you requested), the CA tech advised me to:
"Stop ARCserve ,unload modules nwtape.cdm and nwaspi.cdm .
(Refer link ->

http://support.ca.com/techbases/as61/18022.html

for more information).

Remark the modules in startup.ncf and then reboot the Novell server and then start ARCserve for changes to take effect."

As mentioned in the above link (I think we're all pointing to similar CA links <grin>), ARCserve has perfected their own CANWPA.CDM which takes the place of the Novell drivers.

My tape hardware, for what it's worth, is the HP SureStore DLT Autoloader 718.

To expand on #4, the CA tech asked me to remove SP2 and then apply the QO05996 patch and the QO19341 patch on the ARCserve server and then apply the QO19337 on the remote Netware server with Netware agent on it. I'm going to do that tonight and see what happens tomorrow, since I do happen to have those old patches (no longer available on the CA website).

Another e-mail from CA has suggested that after the first successful backup (after rebooting the server),unload and reload the Netware agent on the remote server and then configure a backup of the remote server. And to check if the second backup job is successful after unloading and reloading the Netware agent.

On that note I will point out one other thing I noticed. When the job hangs, on the remote server (running the Agent), displayed on the ARCserve Tape Server screen:
Active Operations: 2 (usually says 1)
I/O Buffers was 1,966,000 (usually says 0).

We may be narrowing it down to the AGENT, not the ARCserve server. And, to finish on a whiny note, I AM doing CA's bug troubleshooting, but not on their payroll.....

PS: To answer 8arcserve, I'm using GroupWise 6 and my troubles with ARCserve began when I upgraded to version 7. Every patch fixes some problems and creates other problems.

 

[LolaRed]

I got a good backup last night using the QO05996 patch and the QO19341 patch on the ARCserve server and the QO19337 on the remote Netware server with Netware agent on it (I knew I would).

I put SP2 back on and am running a backup tonight using the Agent. Tomorrow I will stop & restart the Agent (not ARCserve) and see if a second backup will succeed.

 

[4gandalf]

8arcserve, I'm having the identical issue with my nwagents.

Could you please send me a the qo28844 patch. My email address is stewartg(at)stlouiscity.com.

Much thanks.

 

[LolaRed]

The ultimate fix (a workaround anyway) is to simply load and unload the agents before each backup job. I use the cron utility to do this daily. My crontab (residing in \\<server name>\sys\etc) file contains the following:

# fields are:
# Min Hr Dat Mo Day Command
#
#recycle nwagent daily at 10:00 am
0 10 * * * unload nwagent
0 10 * * * nwagent

Note that the server where the cron utility runs is the server where the agent resides, NOT the server where ARCserve is installed.

This daily agent recycling keeps the backup job from freezing up. I put this in place in June and notified Computer Associates. Still no real fix from them, but this gets me through....