Contentcenter slow

Hello,

We are running TS 6.5 SP3 on Solaris. Contentcenter is suddenly very slow.
Only operations that read/write the backing store seem to be problematic.

Listing files, expanding the directory tree listing workflows have all slowed.

I've looked in the various log files iwtrace iwserver and servletd.log and can't see anything out of the ordinary.

I've cleared the workflow store, Ran and iwfsshrink and also removed editions so that we only have very recent content. The backing store is located on the same server as the TeamSite install so network speed shouldn't be an issue.

Can anyone else recommend any points to investigate.

Thanks.

Find more posts tagged with

Comments

Adam Stoller

Did you check free disk-space?
Look at the output of iwstat -c
Look at what other processes may be running on your server (like zombied iwperl processes)?

dazzlad

Hi,

We have 16GB free on our backing store partition which is over 50% of the capacity.

I'm not sure what the output to iwstat -c tells me, but it looks like this

Store Status
default Running
newdefault Running
workflow Running

ID Thread User Duration Store Operation
0x12030 0x1f root 0.000 - GetServerStatus

Cache Active Dirty To Purge Available
default 3584 0 0
newdefault 6 0 0
workflow 58 0 0
Total 3648 0 0 26352

We don't have any perl processes running. We've rebooted the entire server a few times and it seems slow as soon as it comes up.

As said, I do think it relates to backing store access as the top navigation of contentcenter loads extremely quickly so there doesn't seem to be an issue with static assests wuch as HTML, CSS and images.....

Thanks.

Adam Stoller

Hi,

We have 16GB free on our backing store partition which is over 50% of the capacity..

What about temp and/or swap space?
Virtual memory?

Have you contacted Support? (Not sure what they can do - but it's reasonable to try)

dazzlad


# swap -s
total: 838104k bytes allocated + 382368k reserved = 1220472k used, 6026016k available
# swap -l
swapfile             dev  swaplo blocks   free
/dev/md/dsk/d1      85,1      16 8395184 8395184

The box has 4GB RAM in it so this all looks reasonable to me?

Thanks.

nipper

How did you decide it was read/write to the backingstore related ? Have you tried using the /iwmnt (say via a find),

Pawan

How do you manage the TeamSite users? Are they domain users? Are you using LDAP?

dazzlad

How did you decide it was read/write to the backingstore related ? Have you tried using the /iwmnt (say via a find),

I've just made http requests for the gifs/js/css separately to the content center servlets and noted that there seems to be a real delay in the servlet response - 2 seconds plus for most iw.ccpro servlets.

I did consider doing a find over iwmnt but i've got nothing to make a comparison to, ot at least no similar hardware (our qa and prod servers are much better kit).

ab_list.rptdesign

dazzlad

How do you manage the TeamSite users? Are they domain users? Are you using LDAP?

We are using local users and groups.

moonkhanna

We are using local users and groups.

Have you checked for any stale locks on your branches.. If not yet, get the script for checking these from Support .. Run this script for each of your branches and it will give you a list of all stale locks..

Unlock then using the unlock CLT.

Stale locks are basically created when non empty directory deletions are submitted..Check with Support..

Second would be to check for any stale jobs.. run iwwfobj on your server and see if you can see any jobs that are not seen in GUI..In case, you have very old jobs, that you believe are needed no more, end those..

Hopefully, above should help ..

prem_pratick

Hi,

You can try the below.

1. Intsall httpAnalyzer and then try to do a round of testing on your CMS.
http://devnet.interwoven.com/forums_vb/showthread.php?t=25941&forumid=11
2. Check if cetain steps are taking more time in "First byte download". If yes then its an application issue. You have to raise a ticket.
3. Implement all the IW recommendations.
4. Check for mod_expired.
5. run iwstat -c in ceratin intervals.
6. Have you installed any patch/software on the server.
7.
Log into CMS .
Replace the URL https://localhost/iw-cc/webdeskpro with https://localhost/iw-cc/command/iw.ping
You would get something as
Interwoven Ping
Ping start mark: 1195484115846
Ping end mark: 1195484116171
Ping round trip time: 325 ms.

8. Do a pathping -n .

9. Do you use ssl.

10. If possible try to use the Keynote automated script. This will help in debugging.

11. Check for the number of files in a dir/folder. How many locked files you are having. No of WF Jobs.

12. In fact delete the workflow directory. and recreate it.

13. Check for your browser settings.

14. Is TS slow all the time or at specific times.

15. Check for the server CPU usage, when you are working on TS.

16. Try to increase the cachesize.

17. Recycle all the IW log files on a weekly basis.

I am telling all the above from my personal exp on Solaris/Windows platform. Hope so this will help you.

Thanks
Prem Pratick.