nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Deleting Files - Disk Still Full - Urgent!

sporkit

Our TeamSite 6 disk is currently full. I did the usual to make some room.

(as root)

iwrmed
iwfsshrink run

I've been watching the iwstat command and something seems wrong. The store status usually says running. Right now it says frozen. Is it possible it could be locked up?


iwstat | head -n 20
Store       Status
default     Out of space - frozen
workflow    Out of space - frozen

ID         Thread  User     Duration  Store    Operation
0x628bf21  0x1b    root     0.000     -        GetServerStatus
0x628bf20  0x13    ahauser  0.014     default  nfs2lookup 112211ne1 (64,402f66bc) fcf5e3/fcf5e0
0x628bf1f  0x12    root     0.015     default  tor_nfs2read 0 (64,3e6698) 3e6698/e7acd9a
0x5df5b9f  0xd     -        4799.687  default  Batch Jobs

Store: default
Batch Jobs (running):
Queued                        User    Job                                                                                                               Object ID
[Thu Aug 16 14:50:25 2012]    root    RemoveDuplicateData (Phase 1)
[Thu Aug 16 14:59:40 2012]    root    Cleanup destroyed edition education042312el (area education042312el)                                              0x00002010000000000e6a6072
[Thu Aug 16 14:59:41 2012]    root    Cleanup destroyed edition InsideSHSMD042312 (area InsideSHSMD042312)                                              0x00002010000000000e6a6c31
[Thu Aug 16 14:59:41 2012]    root    Cleanup destroyed edition aboutus042412el (area aboutus042412el)                                                  0x00002010000000000e6a6e4c

The job area has also been stuck on this output for some time. "Cleanup destroyed edition education042312el".

So my questions are:

* Is it possible that the system could freeze when cleaning itself up because of disk issues?
* Is deleting an edition actually deleting files off the server or just unlinking something?
* If I need to make room fast, what can I delete? Workareas? Staging areas? Editions?

My organization is at a standstill till this is resolved! Thanks

Find more posts tagged with

Comments

nipper

If your store is frozen, the shrink and rm edition will not do anything

How much disk is available ?

run

df -h

change /etc/iw.cfg to allow TS to run until 99% (assuming that is above the available disk)

run iwfreeze --

And then the shrink/em edition should start to do some work

You should also see if there are file (log/temp) that can be removed.

sporkit

It took some time, but I found one branch that had 10-20x more editions than others. I removed old editions from that branch and the system went back to normal quickly.

Deleting editions from smaller branches wasn't getting my where I needed to be and I started panicing.

*Phew! Glad it's solved. Thanks for the addition input.

mart2001

As mention before, you should check the iw.cfg file. Use iwgetlocations to make sure you are looking at the correct location for iw.cfg on unix (It can be in iw-home/etc or /etc).

The iw.cfg file lets you specify if disk space falls below a certain % of the disk or a number in megabytes.

You can also try to clean up some log files. access.log can grow huge if not rotated. servletd_out.log could also grow big, along with iwtrace.log. If you have OD on that machine, you can also clear out some logs in OD also.

Adam Stoller

In addition to the suggestions made by nipper and seconded by mart2001 - I would strongly consider trying to find more disk space to allocate to the location of your backing store.

Deleting editions is okay - but remember that a core piece of the functionality of TeamSite is versioning, and when you start removing editions you start decreasing the versioning that is maintained.

Obviously there are degrees to which this makes sense.

If you have 10 editions on each branch - deleting a single edition is likely to gain back disk space but significantly reduce the value of the backing store as far as versioning is considered.

If you have 1000 editions on each branch - deleting a single edition is likely to not gain back much disk space on a case-by-case basis, but will help maintain a relatively steady-state in terms of disk space utilization (assuming you always delete the oldest edition whenever there are 1000th edition is created) - and will also provide you with a reasonable amount of versioning.

Obviously you can adjust the numbers above by various amounts - but hopefully that gives you and idea. Just deleting editions because you don't have sufficient disk space is not an answer unless the number of editions you're talking about is sufficiently large enough to provide the versioning that you should be maintaining.

kailash1

Plead stop running iwfsshrink. Recommend working with support if you need to run iwfsshrink.

Adam Stoller

Plead stop running iwfsshrink. Recommend working with support if you need to run iwfsshrink.

Based on what?

iwfsck with anything more than the '-d' flag and iwfsfix should be done with Support assistance

iwfsck with the '-d' flag and iwfsshrink have never needed Support hand-holding. and shouldn't now. And if you've ever done an iwmigrate you MUST run iwfsshrink because iwmigrate causes massive amounts of duplicate data which iwfsshrink cleans up.